Desertification is typically associated with arid and semi-arid ecosystems in hot climate zones. However, it also occurs in high rainfall, high latitude systems, such as those in Iceland. The physiognomy of Iceland has changed dramatically since the settlement in the 9^th and 10^th centuries when 15 - 25% of the country had woodland cover.  Today birch cover is only 1 %.  Deforestation is often the initial stage in the land degradation process, followed by a series of surface destabilization phases and culminating in massive soil erosion.  Little is known about the processes driving these catastrophic changes.

The climate of Iceland is maritime, characterized by high annual rainfall (> 1000 mm), cool summers and mild winters where snow cover is shallow and ephemeral and air temperature fluctuate around 0 °C.  Freeze-thaw cycles are thus pronounced and frequent during winter.  Conceptual models propose that deforestation increases the frequency and intensifies the amplitude of soil freeze-thaw cycles and hence increases soil surface movement and destabilization.  If those conceptual models are correct, we should see measurable differences in soil temperatures. To test the temperature fluctuation prediction, we (A) quantified soil temperatures in these contrasting land cover types; and (B) experimentally evaluated the importance of vegetation as a moderator of soil temperature fluctuation using 2 and 10 cm insulation mats to simulate different levels of ground cover. WatchDog data loggers (Spectrum Technologies, Inc.) were used for (A); and soil thermometers (Campbell model 107, 21X datalogger [Campbell Scientific Inc.]) were used for (B).  In both cases, sensors were placed on the soil surface and at 5 cm depth. 

Results/Conclusions

Differences in soil temperatures were observed between the woodland and heathland, and different sward types, but variability was high.  Average temperatures (+/‑SE) were ‑0.21 ±0.34°C and ‑1.30 ±0.41°C for the woodland and the adjacent heathland, respectively, indicating higher temperature trend in the woodland.  Soil temperature increased with increased surface insulation.  The average sward surface temperature was ‑2.42 ±0.40°C, but ‑2.47 ±0.15°C, ‑1.07 ±0.14°C and ‑0.52 ±0.06°C for no sward, 2 and 10 cm insulation, respectively. The results indicate that cryoturbations in woodlands should be less severe than in open areas, and they should decrease with vegetation thermal properties.  Surface stability would thus be expected to be greatest in woodlands or where the sward provided good insulation, and such environments should be more resistant to cryoturbations.  However, the surface microtopography does not reflect that as might have been expected.