Impact of soil moisture–atmosphere coupling on Eur

Authors: E B Jaeger S I Seneviratne

Publish Date: 2010/04/06

Volume: 36, Issue: 9-10, Pages: 1919-1939

PDF Link

Abstract

Processes acting at the interface between the land surface and the atmosphere have a strong impact on the European summer climate particularly during extreme years These processes are to a large extent associated with soil moisture SM This study investigates the role of soil moisture–atmosphere coupling for the European summer climate over the period 1959–2006 using simulations with a regional climate model The focus of this study is set on temperature and precipitation extremes and trends The analysis is based on simulations performed with the regional climate model CLM driven with ECMWF reanalysis and operational analysis data The set of experiments consists of a control simulation CTL with interactive SM and sensitivity experiments with prescribed SM a dry and a wet run to determine the impact of extreme values of SM as well as experiments with lowpassfiltered SM from CTL to quantify the impact of the temporal variability of SM on different time scales Soil moisture–climate interactions are found to have significant effects on temperature extremes in the experiments and impacts on precipitation extremes are also identified Case studies of selected major summer heat waves reveal that the intraseasonal and interannual variability of SM account for 5–30 and 10–40 of the simulated heat wave anomaly respectively For extreme precipitation events on the other hand only the wetday frequency is impacted in the experiments with prescribed soil moisture Simulated trends for the past decades which appear consistent with projected changes for the 21st century are identified to be at least partly linked to SMatmosphere feedbacksClimate extremes have a major societal economical and ecological impact as for instance highlighted by the 2003 summer heat wave and drought in Europe Larsen 2003 Heck et al 2004 Ciais et al 2005 Several recent observational Klein Tank and Können 2003 Schmidli and Frei 2005 Alexander et al 2006 DellaMarta et al 2007 as well as modeling studies Christensen and Christensen 2003 Meehl and Tebaldi 2004 Schär et al 2004 Frei et al 2006 report an increase in frequency and intensity of temperature and precipitation extremes both for the recent past as well as for the coming decadesThe physical mechanisms underlying such changes in extremes of temperature and precipitation may relate to changes in largescale circulation Christensen and Christensen 2003 Meehl and Tebaldi 2004 Pal et al 2004 and/or to changes in smallscale physical processes such as soil moisture–atmosphere interactions Seneviratne et al 2006b Vidale et al 2007Owing to the relevance of extremes these research findings highlight the need for a better understanding of the contributing processes and feedbacks which also implies comparison with observations eg Ek and Holtslag 2004 Jaeger et al 2009 Heat waves are generally caused by quasistationary anticyclonic circulation anomalies Fink et al 2004 Black et al 2004 Meehl and Tebaldi 2004 sea surface temperature SST anomalies Black and Sutton 2006 and/or landatmosphere feedbacks Seneviratne et al 2006b Fischer et al 2007a 2007b whereby the latter can act as an amplifying mechanism Similarly for precipitation variability and heavy precipitation events both circulation patterns Martius et al 2006 and land–atmosphere feedbacks may be relevant eg Beljaars et al 1996 Schär et al 1999 Pal and Eltahir 2002The impact of land–atmosphere coupling on climate is mainly determined by SM limitation on evapotranspiration Seneviratne et al 2010 Since largescale field experiments investigating these effects are not feasible one way of assessing the underlying mechanisms is to run climate model experiments with prescribed SM content eg Koster et al 2004 Seneviratne et al 2006b Rowell and Jones 2006 Fischer et al 2007a Conil et al 2007 This method allows to infer causal relationships regarding the effect of SM on climate since the twoway coupling of the atmosphere and SM is removed and the experiments thus investigate only the oneway effect of SM on the atmosphere whereas the atmosphere has no influence on SM Here this procedure is used to disentangle the effect of SM variability on different time scales as well as to investigate the impact of extreme levels of SM on the current European summer climate To this aim a set of regional climate model RCM experiments are performed using the CLM RCM Sect 21 driven with reanalysis and operational analysis data from ECMWF Thereby the main focus of the present study is on impacts of SM on extremes and trends in temperature and precipitation The analysis is performed for the summer season when oceanic impacts on climate are small compared to SM impacts over midlatitudinal land areas eg Koster and Suarez 1995One can distinguish three different approaches for the analysis of climate extremes A first group considers directly the probability density functions PDFs of the investigated variables generally temperature or precipitation Alexander et al 2006 Perkins et al 2007 and thereby focuses on their tail behaviour Since most statistical distribution functions do not well describe the tail behaviour of the underlying data a second group of studies uses techniques of extreme value theory EVT that provide special distribution functions for extremes The study of Frei et al 2006 for instance uses EVT to assess the future change of precipitation extremes in Europe based on a set of RCM experiments from the EUproject PRUDENCE http//prudencedmidk Other modeling studies use EVT to assess changes of temperature extremes eg Zwiers and Kharin 1998 Kharin and Zwiers 2000 There are also several observational studies assessing changes in temperature extremes using EVT which report an increase at least in the location some also in the shape of the used extreme value distribution eg Laurent and Parey 2007 DellaMarta et al 2007 Brown et al 2008 Finally a third group of studies uses so called climate extreme indices to capture a variety of aspects of climate extremes both in models eg Frei et al 2006 Fischer et al 2007a Kjellström et al 2007 and observations eg Klein Tank et al 2002 Schmidli and Frei 2005 DellaMarta et al 2007 Wet dry hot or cold events can be extreme in terms of frequency duration or intensity and these aspects cannot be investigated from the analysis of temperature and precipitation PDFs only As an example the EUFP6 project CECILIA http//wwwceciliaeuorg/ established a list of more than 130 indices characterizing temperature and precipitation statisticsBeside the analysis of the role of SM for climate extremes we also assess in this study the possible impact of SM on climate trends The investigation of trends and their relation to possible changes in drivers or feedback processes has received increasing interest in the climate community due to climate change For climate extreme indices the analysis of trends is mainly performed using either parametric methods eg regression models Klein Tank and Können 2003 Schmidli and Frei 2005 or nonstationary extreme value analysis eg Kharin and Zwiers 2005 or nonparametric methods eg robust slope estimator TheilSen Alexander et al 2006 or digital filters Tebaldi et al 2006

Keywords: