Research Projects
Aerosol-Cloud Semi-Direct Effects In addition to affecting cloud microphysical properties, aerosols--in particular those that absorb solar radiation--also have a thermodynamic effect on clouds. By absorbing solar radiation and heating the atmosphere, ambient relative humidity (RH) is reduced, encouraging cloud evaporation. This leads to an increase in incident solar radiation at the surface, causing additional surface warming. This has been called the semi-direct effect (SDE) (e.g., Hansen et al., 1997; Ackerman et al., 2000) and acts to partially offset the cooling of aerosol indirect effects. Although current global SDE estimates are small compared to indirect effects, the SDE remains largely unquantified and highly uncertain, with the few relevant studies leading to different conclusions on the magnitude and sign of the SDE. Recent GCM results with NCAR’s Community Atmosphere Model (CAM3) forced with satellite-based aerosols (Chung et al., 2005) yielded a global semi-direct effect significantly larger than prior estimates. The dominant effect was a reduction in mid-level cloud, which warmed the surface and negated up to 100% of the impact of aerosol direct effects on surface fluxes. More importantly, the SDE contributed to climate change in unexpected ways, most notably by a Northern Hemisphere land-sea warming contrast. The figure below shows the change in low cloud (CLOW) for June-July-August due to anthropogenic aerosols (symbols represent a response significant at the 90% confidence level or greater). Over land, low clouds decreased due to RH reductions; over the sea, however, low clouds increased due to increases in lower-tropospheric stability. Also shown is the corresponding top-of-the-atmosphere semi-direct effect. In the NH extratropics, the SDE over land was +2.54 W m−2, compared to -0.56 W m−2 over the sea; clearly, these radiative flux changes are largely driven by CLOW and act to warm the land, but cool the ocean.  Role of non-CO2 Anthropogenic Forcings in Tropical Expansion Recent observational analyses show the tropics have widened over the last several decades. Estimates range from 2-5° of latitude since 1979 (Seidel et al., 2008) and are based on several metrics. Although tropical expansion has many important societal implications, particularly from a hydrological perspective, the mechanisms driving the expansion are not well known. Moreover, state-of-the-art GCMs significantly underestimate the rate of recent expansion. Using the Intergovernmental Panel on Climate Change (IPCC) Coupled Model Intercomparison Project, Phase 3 (CMIP3), Johanson and Fu (2009) found the largest CMIP3 tropical widening trends are ~1/5 of the observed widening. This significant underestimation existed across five scenarios, as well as three separate definitions of Hadley cell width, including dynamical and hydrological definitions.  Idealized tropospheric heating experiments with the CAM3 GCM show that mid-latitude heating results in significant poleward displacement of the tropospheric jets, whereas high-latitude heating shifts the jets equatorward. Warming the tropical troposphere primarily strengthens the jets, with minimal zonal displacement. This is illustrated in the figure to the left, which shows the zonal annual mean temperature and zonal wind response to lower tropospheric heating at high-, mid- and low-latitudes. An increase (decrease) in U on the poleward (equatorward) flank of the jet (black contour lines) is indicative of poleward jet displacement, as illustrated by the mid-latitude heating experiment. Such jets shifts are consistent with thermal wind balance and the corresponding displacement of the latitude of the maximum meridional tropospheric temperature gradient. Over 70% of the annual mean jet displacements in a suite of CAM3 global warming experiments, as well as in CMIP3 2xCO2 equilibrium experiments, is accounted for by an ‟Expansion Index”, which compares mid- and high-latitude tropospheric warming amplification. The following figure shows the tropospheric poleward jet displacement versus the Expansion Index for the CAM global warming experiments and 12 CMIP3 2xCO2 equilibrium experiments.  These findings suggest recently observed tropical expansion is partially driven by mid-latitude heating. Deficient absorbing aerosol and tropospheric ozone forcing--which primarily warm the NH mid-latitudes--may have contributed to model underestimation of recent tropical expansion. Effects of Eurasian Snow Cover on the Dominant Mode of NH Wintertime Atmospheric Variability Recent observational studies have revealed a relationship between autumnal snow cover over Eurasia (EA) and the subsequent winter’s Arctic Oscillation (AO). The figure below (after Cohen et al., 2007) shows the observed correlation between October EA snow cover and daily polar-cap geopotential height (ZPC), which is used as a representative measure of the AO. Anomalous October EA snow cover is associated with positive December-January ZPC anomalies (which are associated with the negative phase of the AO). The figure also shows a conceptual model of the proposed snow-AO mechanism. Anomalous EA snow cover leads to strong diabatic surface cooling, which increases the flux of wave activity from the troposphere. This energy is absorbed in the stratosphere, which leads to higher geopotential heights and a weaker polar vortex. These anomalies then propagate down from the stratosphere, culminating in a strong negative phase of the AO at the surface.   Idealized modeling studies also support the snow-AO relationship. The figure below shows the time-pressure cross section of the ensemble mean polar-cap (60-90N) geopotential height response (ZPC), and the vertical component of the wave activity flux (WAFZ), based on idealized CAM3 experiments forced with prescribed, anomalously high EA snow albedo. Similar to other non-CAM GCM studies using prescribed snow mass (Fletcher et al., 2009), anomalous EA snow albedo produces a wave activity pulse that propagates into the stratosphere, culminating in a negative phase AO-like surface response (as represented by positive ZPC anomalies).  When GCMs are allowed to freely simulate snow cover, however, they fail to capture the snow-AO relationship (Hardiman et al., 2008). CAM3 is no exception, as transient simulations with prognostic snow cover yield a snow-AO relationship inconsistent with observations, and dissimilar AO trends. However, as with most GCMs, CAM3’s EA snow cover, and its interannual variability, are significantly underestimated (October EA snow cover is underestimated by 40% and its standard deviation by 85%). When the albedo effects of snow cover are prescribed in CAM3 (CAM PS) using NOAA satellite-based snow cover fraction data, a snow-AO relationship similar to observations develops. Furthermore, the late 20th century increase in the AO, and particularly the recent decrease, is reproduced by CAM PS. The figure below shows the 1972-2006 correlation between September-October-November EA snow cover and August-March polar-cap geopotential height (ZPC) for CAM and CAM PS (compare to the Cohen et al., 2007 plot above, based on observations).   Snow-Hydrological Feedback and Quasi-biennial N/AO Persistence  Variations in the Arctic Oscillation (AO), and its regional manifestation the North Atlantic Oscillation (NAO), generate much of the non-seasonal variability in winter climate over the Northern Hemisphere (NH) mid-latitudes. Despite being an internal mode of the atmosphere, the N/AO exhibits a slightly red spectrum, varying on quasi-biennial (2-3 years) and quasi-decadal times scales. Such low-frequency variability is likely due to coupling of the atmosphere to boundary conditions and/or external forcings. The figure above shows that Eurasian (EA) snow cover--particularly over eastern Siberia (ESB)--exhibits quasi-biennial persistence similar to the N/AO, with moderately significant, positive autocorrelations at lag-1 and -2. Moreover, the snow-AO mechanism operates on quasi-biennial time scales, with fall ESB snow cover significantly related to vertically propagating Rossby wave activity, and the N/AO, for the next 2-3 winters. The figure below shows 1972-2007 lag correlations between OCT ESB snow cover and the monthly NAO and AO for the subsequent five winters (lag correlations are plotted from October at lag-0 through April at lag-5, with the ith lag denoted with “+i”). Positive ESB snow anomalies are significantly correlated with the negative phase of the N/AO for the subsequent winter (J+1), as well as the next 2 winters (J+2 and J+3).  Although anomalously high autumnal EA snow generally persists to the subsequent late spring/early summer, the anomaly does not persist through August, when nearly all (except over the Tibetan Plateau) of the October snow anomaly has melted. Because the snow anomaly does not persist through the summer, yet reappears the following fall, a mechanism must exist that provides the climate system with memory of the October snow anomaly. The figure below, based on the Variable Infiltration Capacity (VIC) land surface model driven by the Global Land Data Assimilation System (GLDAS), suggests the interseasonal carryover of ESB snow is related to snow moisture anomalies and an evaporation-convection feedback (Yeh et al., 1983; Dirmeyer and Brubaker, 2007). The figure shows the second empirical orthogonal function (accounting for 8% of the joint variance) based on a simultaneous PCA of 7 fields, including winter (JFM) snow water equivalent (SWE), summer (JJA) rain rate (RR), evapotranspiration (ET), and soil water (SW), October snow rate (SR) and October NOAA snow cover area (SCA), as well the prior October’s SCA. Such an analysis allows the identification of simultaneous patterns of variation between multiple fields.  Over eastern Siberia, this mode is associated with similar signed variations of all seven variables. This includes a strong ESB relationship between JFM SWE, JJA SW, JJA ET, and JJA RR. This supports the notion high winter ESB snow mass is associated with high ESB summertime soil moisture, which is sustained through an evaporation-convection feedback (high ET and RR). Although the mode is weakly associated with positive ESB NOAA SCA anomalies in October, the relationship with OCT snow rate is stronger. Moreover, the bottom panel shows that this mode is associated with high ESB SCA in both the current and previous OCT, and this area closely coincides with the area of ESB OCT SCA persistence.