Session: Last Millennium & Past2K
Author: Fredrik Charpentier Ljungqvist / email@example.com / Stockholm University
An increasing risk of droughts in some regions, with potentially severe consequences for agriculture and economy, entails a major challenge associated with ongoing and future global warming. However, climate model simulations do not show consistent projections as to hydroclimate changes with global warming on regional scales. We also have increasing evidence from around the world that the relationship between temperature and drought is highly timescale-dependent – thus the relationship seen in instrumental measurements over shorter time-scales might not hold true for longer time-scales.
Therefore, a new PAGES2k project will study how the variations in the spatio-temporal distribution of droughts during past warm and cold periods in Europe can provide tentative information for future changes in European droughts associated with global warming. In light of recent progress in developing high-resolution European climate reconstructions it is now possible for the first time to assess the co-variability between summer temperature and drought frequency and severity over the past millennium.
To realize this goal, we assess the co-variability between an updated version the Old World Drought Atlas, providing a spatially revolved tree-ring based gridded summer drought index for the European-Mediterranean area extending back two millennia, and a spatially resolved summer temperature reconstruction from tree-ring and historical documentary from 850 to 2003 CE. Additionally, we compare the high-resolution summer temperature and soil moisture simulations from the CCSM4 and MPI-ESM-P models over the same time period with the proxy-derived results. We also compare the co-variability between summer temperature and drought in the CRU TS 3.21 instrumental data for 1901–2012. We perform the comparison of the co-variability by: 1) cross-correlation calculations between gridded instrumental, proxy, and model fields, 2) sign tests of agreement between gridded instrumental, proxy, and model fields, 3) analysing the distribution of correlations in the various data series, 4) performing cross-spectral analyses of the various data series, and 5) conducting cluster analyses of the various data series. Preliminary results suggest that the co-variability between temperature and drought indeed depends on the time-scales chosen and spatial patterns of co-variability are more complex in the proxy-derived reconstructions than in the model simulations.