by Gijs van Lith
Arctic temperatures have been rising more than the global average. This greater rise is due to the many feedback systems that react on each other, thereby amplifying the effects; arctic amplification. The retreat of permafrost has lead to a change in vegetation activity. This change has an amplifying effect on the surface temperature change. A factor that might influence this relationship of surface temperature and vegetation activity is surface water. The study area of this research are the low lying areas of Northern Siberia. This has a thermokarst land surface, and is therefore dominated by water bodies. Due to the characteristics of the region and the changes in that are occurring in the arctic, the area covered by water is very volatile.This research investigates if sub pixel surface water changes have an effect on the spatiotemporal relation of surface temperature and vegetation activity. The period considered is 2000 – 2011. These variables were translated into indices measurable by satellite mounted sensors. The Normalised Difference Vegetation Index (NDVI) is used to measure vegetation activity, and the Summer Warmth Index (SWI) is used to measure surface temperature. These indices were calculated from MODIS data. Different water indices were tested to see which was best in detecting the size of the area covered by surface water. This was done by using a multi scale validation with images of different resolutions.
To find out if the amount of surface water has an influence on the relationship of surface temperature and vegetation activity, linear regressions were performed on time-averaged images index rasters. To find out if there were changes over time of the surface water, surface temperature, and vegetation activity, trends were set up using the Mann Kendall trend test. To discover the relationships between the surface water dynamics, surface temperature, and vegetation activity, the Mann Kendall rank test was performed on detrended time series of the indices. Detrending was done using the Lowess smoother.
None of the water indices gave a good estimation of the area covered by surface water. However, the Modified Difference Normalised Water Index (MDNWI) of Xu (2006) gave the best indication of the presence of water. This index was further used to represent the amount of surface water in this research. The amount of surface water present has a negative effect on both surface temperature and vegetation activity. It also has a negative effect on the relationship of these two variables, as there are larger errors at higher surface water levels. Northern Siberia shows a slight increase of surface water, an increase in surface temperature, and a slight decrease in vegetation activity over the period considered.
Surface water dynamics have a small effect on the dynamics of surface temperature and vegetation activity. At more profound water dynamics the dynamics in surface temperature and vegetation activity are also more profound. The water dynamics are overall positive in the region, and the dynamics of vegetation activity and surface temperature are mostly negative. More profound water dynamics cause for a better relation of the dynamics of surface temperature and vegetation activity. Positive water dynamics give a stronger relation of surface temperature and vegetation activity than negative dynamics. Landscape characteristics seem to have an influence of the relationship of surface water dynamics and dynamics of surface temperature and vegetation activity.
This research shows that surface water dynamics have an effect on the arctic amplification. To get a better insight in the future levels of surface water, changes in permafrost should be researched, as this seems to have a large impact on the presence of surface water.
Keywords: MODIS; NDVI; NDWI; MNDWI; SWI; Northern Siberia; arctic; arctic amplification; global warming; feedback mechanisms; time series.