ATS 150 Global Climate Change Climate Feedback, & Sensitivity Planetary Energy Balance Climate Sensitivity How Many Degrees of Warming per Watt/m2 of Heating? Energy In = Energy Out Please read Chapter 7 S(1−α)π R2 = 4π R2σT 4 in Archer Textbook T ≈ −18o C But the observed Ts is about 15° C Earth’s Climate Climate Forcing, as a “Black Box” Response, and Sensitivity S0 (1−α) F = TS T 4 Climate ΔF Climate Δ S 240 15 °C + 1 ? °C W m-2 System W m-2 System Surface Absorbed Forcing Response: Sunshine Temperature (change in (Change in absorbed Surface In Out sunshine) Temperature) Scott Denning CSU Atmospheric Science 1 ATS 150 Global Climate Change Climate Feedback, & Sensitivity Climate Forcing, Baseline Climate Response, and Sensitivity Sensitivity “Let’s do the math …” S (1−α) F = 0 = σT 4 Forcing Forcing 4 Response: Response: Temperature) Temperature) dF 3 3 2 1 (Change in Surface in Surface (Change − − (change in sunshine) (change “Let’s do the math …” = 4σT = 4σ (255K) = 3.8W m K dT S (1−α)πr2 = 4πr2σT 4 0 1 0.266 K 4 ΔT = −2 −1 ΔF = −2 ΔF S0 (1−α) = 4σT 3.8W m K W m S (1−α) F = 0 = σT 4 4 A 1 W m-2 change in absorbed sunshine produces Solve for ΔF that produces a given ΔT about a 0.27 °C change in Earth’s temperature Climate Feedback Processes Ice Albedo Feedback • Positive Feedbacks Positive feedback (amplify changes) – Water vapor Δ high cloud – Ice-albedo Lapse Δ – High clouds Δ low cloud F T • Negative feedbacks Δ Δ S vapor Δ (damp changes) • Radiative forcing melts snow and ice – Lapse rate • Darker surface absorbs more radiation albedo Δ – Low clouds • Amplifies warming or cooling Scott Denning CSU Atmospheric Science 2 ATS 150 Global Climate Change Climate Feedback, & Sensitivity Water Vapor Feedback Lapse Rate Feedback Positive feedback • Radiative forcing warms surface • Warmer surface evaporates more water • Greenhouse effect depends on emission to space from Positive OR negative feedback! feedback! OR negative Positive higher (colder) levels of the atmosphere • Warmer air can “hold more water” • If radiative forcing produces increased vertical mixing • Increased water vapor (GHG) absorbs by convection, then more heat is mixed to higher levels more outgoing radiation, amplifying • Warm air aloft emits more radiation to space, warming compensating for original forcing Cloud Feedbacks Estimating Total Climate Sensitivity 1. Paleoclimate analogs: how much has climate changed in the past when forcing of known strength was applied? – Advantage: all feedbacks included – Disadvantage: hard to know exactly how much forcing & global temperature response 2. Calculation from physical principles including feedback processes Positive OR negative feedback! feedback! OR negative Positive (complex global climate models) • Additional water vapor makes more clouds – Advantage: Physical insight • Low clouds cool, but high clouds warm – Disadvantage: “All models are wrong …” Scott Denning CSU Atmospheric Science 3 ATS 150 Global Climate Change Climate Feedback, & Sensitivity Learning from the Past Ice Age World 1. Geologic past (100’s of millions of years) 2. Deglaciation analog (18,000 years ago to preindustrial time) 3. Last Millennium analog (Medieval Warm Period to Little Ice Age) 4. Modern Climate Record (20th Century changes) The further back we go, the less data we have to work with. High albedo Using modern data, we have only brief transients to study. Low CO2 CO2 and the Ice Ages LGM Climate Forcing • Over the past 420,000 years atmospheric CO2 has varied between 180 and 280 ppm, beating in CO2 time with the last four 300 370 ppm in 2000 glacial cycles 275 250 ice ice ice 225 ice 200 Vostok (400k yr) Ice Core data (Petit et al, 1999) 175 -400000 -300000 -200000 -100000 0 Year Source: Hansen and Sato (2011) Scott Denning CSU Atmospheric Science 4 ATS 150 Global Climate Change Climate Feedback, & Sensitivity Our Variable Star Cycle of Solar Variability Climate forcing = ΔS(1-α)/4 ~ 2 W m-2 x 0.7 / 4 = 0.35 W m-2 11-year magnetic cycle active sun 0.15% brighter BOOM! Volcanic Aerosol • Volcanoes release huge amounts of SO2 gas and heat • SO2 oxidizes to SO4 particles (“aerosol”) and penetrates to stratosphere • SO4 aerosol scatters solar radiation back to space Mt. Pinatubo, 1991 Scott Denning CSU Atmospheric Science 5 ATS 150 Global Climate Change Climate Feedback, & Sensitivity Particles “Aerosol” Particles in Troposphere from air pollution Tropospheric Aerosol Reconstructed Climate Enhance Cloud Albedo Forcing Since 1000 AD • Ship tracks off west coast 1816 1991 0.5 • Cloud droplets Pinatubo condense on Tambora “year w/out tiny particles 1.0 W m-2 a summer” • Makes more/ smaller cloud drops • Clouds are brighter (higher albedo ) Scott Denning CSU Atmospheric Science 6 ATS 150 Global Climate Change Climate Feedback, & Sensitivity R EPORTS Because the concentration of water vapor satellite-observed radiances, thereby eliminat- The above simulations show that the mod- mass decreases roughly exponentially with ing errors associated with the retrieval process el atmosphere dries in response to the radia- height, changes in the total column water vapor (36). Figure 3 compares the satellite-observed tively induced cooling in agreement with the are dominated by the response of water vapor in equivalent blackbody temperatures at 6.7 ␮m observations. But to what extent does this the lower troposphere. Yet, water vapor in the (T6.7)fromtheTOVSinstrument(37)with drying amplify the cooling? To answer this upper troposphere has a disproportionatelyThethose Past computed offline from2000 the model’s tem- Yearsquestion, we repeat our experiments using a Second Millennium Analog large effect on the outgoing LW radiation (34) perature and moisture profiles (38). Under clear configuration of the model in which water and, consequently, on climate sensitivity (5). skies, the 6.7-␮mchannelisprimarilysensitive vapor feedback has been artificially sup- Therefore, we also examine the response of the to changes in relative humidity averaged over a pressed by removing the LW component of upper tropospheric water vapor between 300 deep layer of the upper troposphere (roughly the feedback loop (26, 42). Once again, three • Cooling from Medieval and 500 hPa. Because retrievals of upper tro- 200 to 500 hPa) (39). Thus, if the water vapor pairs of integrations are performed, where pospheric water vapor are less reliable than mass in the upper troposphere decreases by each pair consists of a control simulation and Warm Period to those for the total column, we include both conserving relative humidity as the atmosphere a Mount Pinatubo simulation using the same Little Ice Age ~ 0.8 K NVAP (25)andTIROS(TelevisionInfrared cools, only a small perturbation to T6.7 would aerosol forcing as before. Observation Satellite) Operational Vertical be expected. Figure 4 compares the Mount Pinatubo– Sounder (TOVS) products (26, 35)inourcom- Figure 3 confirms that both the observations induced cooling from both the “standard” parison (Fig. 2, bottom). and model simulations yield only a modest re- (i.e., with water vapor feedback) and “no -2 Both sets of observations indicate a drying duction in T6.7.Infact,themodel-simulated water vapor feedback” configurations of the • Solar forcing ~ 1.0 W m of the upper troposphere after the Mount Pina- anomalies are nearly identical to those obtained model. In contrast to the standard model, the (somewhat complicated tubo eruption, although TOVS suggests a if one repeats the calculation of T6.7 under the model without water vapor feedback is un- somewhat larger drying than does NVAP. The assumption of a constant relative humidity able to reproduce the observed cooling. Over by volcanic forcing) model simulations show similar reductions dur- change in the model’s water vapor field [shown the period of June 1991 to December0.8 K 1995, ing the first half of the record, where the cool- as a green curve in Fig. 3 (40)]. In contrast, the standard model predicts an average global ing is greatest. After mid-1993, the NVAP consider the anomalies in T6.7 that would result cooling of 0.31 K, which compares favorably anomalies rapidly return to pre-eruption levels, if there was no decrease in water vapor mass in with the observed cooling of 0.30 K [0.33 Ϯ Total climate whereas the model simulations and the TOVS the model’s upper troposphere [shown as a red 0.03K, with ENSO signal removed (43)]. retrievals show a more gradual return. Howev- curve in Fig. 3 (41)]. In this case, the T6.7 would Without water vapor feedback, the model- er, given the difficulty in retrieving such small decrease by ϳ0.8 K, more than twice what was predicted cooling is only 0.19 K. Thus, feed- sensitivity concentrations of water vapor, it is unclear observed, due to a reduction in Planck emission back from water vapor amplifies the magni- whether this discrepancy reflects a deficiency in (cooling) without a compensating reduction in tude of global cooling by ϳ60%, which is in (including feedback) the model or in the inversion algorithms used atmospheric opacity (drying). Thus, without a good agreement with the amplification pre- -2 by NVAP. nearly constant-relative-humidity drying of the dicted by climate models in response to a 0.8 K per (W m ) 1826 This uncertainty can be avoided by compar- upper troposphere, the model would be unable doubling of CO2 (3) and with that derived ing the model simulations directly with the to reproduce the observed record of T6.7. from idealized calculations using a constant relative humidity approximation (5). To reproduce the observed temperature Fig. 3. Comparison of record after the eruption of Mount Pinatubo, the observed (black) the model requires a strong positive feed- and GCM-simulated (blue) changes in glo- back, equivalent in magnitude to that predict- bal-mean (90°N–90°S) ed for water vapor.