CONTRAILS CLIMATE STUDIES Patrick Minnis NASA Langley Research

CONTRAILS & CLIMATE STUDIES Patrick Minnis NASA Langley Research Center Hampton VA, USA 30 October 2003 Langley Research Center

MOTIVATION • Air traffic increasing 2 - 5%/year over the globe • Ice supersaturation exists 10 -20% of the time at flight altitude • Aircraft produce persistent contrails => cirrus aviaticus • Cirrus clouds affect radiation budget, possibly water budget • Aircraft exhaust might affect microphysics of extant cirrus • Contrail/cirrus impact least certain effect of aircraft on climate Can contrails have an effect large enough for concern? - Mitigation efforts by aircraft industry (new technology) - Mitigation efforts by air traffic control (new routing) Langley Research Center

Aircraft exhaust short circuits natural cirrus formation - high humidities normally needed to make cirrus (C) - cirrus can exist at lower humidities (B) but need formation boost - no cirrus for RHI < 100% (A) T < -39°C Langley Research Center

GOES-8 IR LOOP FOR NOVEMBER 18, 2001, 1015 - 2115 UTC Langley Research Center

11 -12 µm temperature difference from 1 -km satellite data 24 October 2003; Okla, Ark, Kan, Missouri NOAA-15 1250 UTC NOAA-17 1738 UTC Langley Research Center

11 -12 µm temperature difference from 1 -km satellite data 24 October 2003, Okla, Ark, Kan, Missouri <= Terra MODIS 2025 UTC NOAA-12 IR 2251 UTC Contrails have become cirrus clouds Langley Research Center

11 -12 µm temperature difference from 1 -km satellite data 24 October 2003, Midwest <= NOAA-15, 1250 UTC NOAA-12, 2111 UTC => More contrail cirrus Langley Research Center

11 -12 µm temperature difference from 1 -km satellite data 24 October 2003, Other areas TEXAS Idaho CA coast Pacific NW Langley Research Center

CONTRAILS • Ubiquitous feature of our skies - increase cirrus coverage over areas with air traffic • Can affect climate by altering - Radiation budget (warming, cooling) - Changing moisture budget of upper troposphere? Langley Research Center

THEORETICAL RADIATIVE EFFECTS OF CONTRAILS Langley Research Center

MEAN DAILY HEATING RATES DUE TO CONTRAILS t = 0. 3 t Langley Research Center

CONTRAIL UNKNOWNS • contrail-cirrus coverage - geographical & temporal - now & future (requires modeling) • microphysics: optical depth, particle size - t mean between 0. 1 and 0. 4, varies between 0. 01 & 2 - De changes over life cycle (5 -100 µm) • radiative forcing - depends on when and where it occurs • vertical spreading - dries the UT? Langley Research Center

APPROACH Can it be significant? • Estimate lower & upper bounds of current contrail-cirrus impact - use empirical-theoretical estimates in RTM - relate cirrus change to air traffic Can we accurately predict it? • Develop climatology of contrail coverage, frequency, microphysics, radiative forcing - surface & satellite observations • Relate contrail observations to meteorological conditions - develop empirical-theoretical models to predict contrail coverage & properties Langley Research Center

BACKGROUND • European studies - regional coverage from linear features in AVHRR imagery - tuned global coverage from ECHAMP/ECMWF output (Sausen et al. 98) - recently estimated t = 0. 11 from AVHRR, similar from GCM - GCM simulations of CRF (Contrail Radiative Forcing) • US studies - La. RC estimated t = 0. 30 from AVHRR & GOES data over US - NASA GISS GCM simulation of CRF (Rind et al. 2000) - La. RC simulation of CRF from European tuned output/ISCCP/ERBE - lower bound (Minnis et al. 99) Langley Research Center

La. RC Contrail Minimum Radiative Forcing Estimates => Global contrail forcing: FSW = -0. 003 to -0. 012 Wm-2, FLW = 0. 011 to 0. 033 Wm-2 Fnet = 0. 008 - 0. 020 Wm-2 ; 0. 017 Wm-2 for t = 0. 3 European estimate 0. 003 Wm-2 for t = 0. 15 Greater over areas with heavy air traffic! Langley Research Center

Current Estimates of contrail radiative forcing Minnis et al. 1999, GRL Langley Research Center

La. RC Contrail Maximum Radiative Forcing Estimates • Estimate change in cirrus cloudiness due to air traffic - primary : sfc obs 1971 -1995 • Repeat CRF calculations with cirrus change estimate - assume linear scaling with coverage, t = 0. 15 - 0. 25 • Use GCM conversion factors to estimate temperature changes Rind et al. (2000): = > 0. 025 Wm-2 for t = 0. 25 New range of global radiative forcing = >0. 006 - 0. 025 Wm-2 Langley Research Center

To estimate upper bound contrail radiative forcing: • Measure trends in cirrus where contrails form & do not form (air traffic patterns) • Estimate impact of relative humidity • Estimate cirrus change for no humidity change - no trend over USA Study in "Contrails, Cirrus, and Climate, " Minnis et al. , 2003, accepted J. Climate Langley Research Center

EVIDENCE FOR CHANGE IN CIRRUS CLOUDINESS DUE TO CONTRAILS IS PILING UP! Langley Research Center

Trends in cirrus cover ( SFC OBS, 15+ yrs) & RH(300 h. PA), 1971 -1995 Cirrus trend 1992 contrail cover Langley Research Center Cirrus trend, conf level 90% RH from Minnis et al. 2004

Langley Research Center

Cirrus & contrail seasonal trends Surface cirrus: 71 -95 Satellite contrail coverage: 1990 s (Mannstein et al. 1998, Palikonda et al. 2002) Satellite contrail frequencies: 1993 -94 & 98 -99 (Minnis et al. 2002) USA Europe -Contrails consistent with cirrus trend over USA, not Europe Langley Research Center

Table 2. Contrails, mean cirrus cover, and cloudiness trends (%/decade) over air traffic regions from surface (CC) and ISCCP (CCI) data. The numbers in parentheses indicate the interannual variability in CC. The 1971 -95 trends in CC are all significant at the 99% confidence level, except over WEUR where no trend is apparent. Region 1992 ECO N (%) Mean CC (%) 1971 -95 CC Trend 1971 -95 Total Cloud Trend 1971 -95 CC Trend, 1983 -95 CCI Trend, 1983 -95 WASIA 0. 08 36. 2 (1. 0) -0. 9 -0. 7 -2. 0 -2. 1 EUR 0. 60 18. 5 (1. 3) -1. 2 -0. 4 0. 0 WEUR 1. 52 19. 8 (1. 4) 0. 0 -0. 7 1. 8 0. 9 USA 1. 75 29. 2 (1. 1) 1. 0 0. 5 0. 3 2. 3 LOR 0. 09 24. 5 (0. 5) -1. 6 -1. 4 -1. 5 -0. 6 NA 0. 32 15. 3 (0. 7) 0. 7 0. 0 0. 3 0. 2 NP 0. 16 15. 7 (0. 8) 0. 9 0. 8 1. 6 -0. 4 0. 13 14. 4 (0. 6) Langley Research Center 0. 7 1. 2 0. 8 0. 1 OOR Minnis et al. 2004, J. Climate

Zerefos et al. 2003, JGR Cirrus coverage trends more positive over areas of heavy air traffic in a given region Based on ISCCP data Langley Research Center

Mannstein et al. , AAC, 2003 Over Europe, cirrus coverage, especially thin cirrus, coverage highly dependent on air traffic Based on Meteosat imagery linear contrail coverage over Europe only 0. 3% cirrus delta = 3% Contrail spreading a factor of 10! Langley Research Center

INTERIM SUMMARY • Cirrus coverage is increasing over USA (consistent w/ seasonal contrail frequency steady over Europe (inconsistent) decreasing over western Asia, but not in areas of heavy air traffic decreasing most over other land areas • Cirrus is increasing over ocean (not many obs in pristine areas) Is increase due to air traffic or weather changes? - Zerefos and Mannstein results suggest the former - Minnis et al. (2004) agree Langley Research Center

Langley Research Center

ESTIMATION OF TEMPERATURE CHANGE OVER USA DUE TO CONTRAIL CIRRUS BASED ON GCM STUDY & CIRRUS OBS Langley Research Center Minnis et al. 2004, J. Climate

IMPACT OF CIRRUS TREND • Contrail cirrus can account for all of observed warming over USA between 1975 & 1994 - Ozone impact not included! Langley Research Center

CHANGES IN ATMOSPHERIC TEMPERATURE, 1979 -1997 200 -850 mb FROM SATELLITE DATA AND ESTIMATED CONTRAIL RADIATIVE FORCING data from J. R. Christy Minnis et al. GRL (1999) present study con < 0. 2% Langley Research Center

ESTIMATES OF ALL AVIATION FORCING WITHOUT CONTRAIL SPREADING - IPCC (1999) Langley Research Center

Contrails are a problem, what can we do? BETTER UNDERSTANDING & PREDICTION • IMPROVED OBSERVATIONS OF LINEAR CONTRAILS & SPREADING - provides data for developing & validating models • RELATIONSHIP BETWEEN AIR TRAFFIC, CONDITIONS, & CONTRAILS - gives basis for parameterizing cirrus aviaticus • PREDICT WHEN & WHERE CONTRAILS WILL FORM Langley Research Center

Linear Contrail Climatology Automated Contrail Detection NOAA-12 AVHRR, April 1997 VA NC 10. 8 -µm image detected contrails methodology from Mannstein et al. 1999 Langley Research Center

LINEAR CONTRAIL COVERAGE DURING 2001 FROM AVHRR Palikonda et al. 2003) 730 AM APRIL DECEMBER Langley Research Center 230 PM

CONTRAIL DETECTION FROM SATELLITES • Very sensitive to particular imaging instrument response - need careful tuning of technique • Sometimes mistakes cirrus streaks as contrails - need error budget • Sometimes misses larger contrails - need more error budget This work is underway! Langley Research Center

Optical depths nearly identical for both NOAA 15 & 16 Langley Research Center

Comparison of contrail coverage (%) USA Time Sausen et al. 98 Palikonda et al. 98 1993 -94 present (NOAA-16, 15) 2001 Dec 1. 6 2. 1 (Dec) 0. 8 (0. 9) Apr 2. 0 0. 7 (1. 3) Jul 0. 5 1. 3 0. 3 (1. 1) Oct 1. 9 0. 8 (1. 0) ---------Sausen et al. 98, Global Annual Langley Research Center 0. 087

Comparison of contrail properties Source Time OD NCLRF (Wm-2) Minnis et al. 98, USA Apr 0. 30 (theoretical) Annual 0. 30 Palikonda et al. 98 Apr (0. 27) 12. 4 (14. 2) Jul (0. 30) 16. 0 (22. 3) Oct (0. 27) (10. 4)) Dec (0. 27) 11 (12. 0) Annual 0. 11 Minnis et al. 99, Global N 14, 93 -94, USA ( N 15&16, 01, USA) Meyer et al. 02, Europe (NOAA 14, 95 -97) Langley Research Center 27 14

CONTRAIL PREDICTION • To relate contrails to the conditions, we have acquired a database of flight tracks for commercial air traffic over USA for 3 years - Garber et al. 2003 (NASA RP in review) • Use NCEP Rapid Update Cycle (RUC-2) experimental product to predict contrail occurrence - realtime USA contrail predictor online new RUC data not very good for contrails • Compare with satellite data Langley Research Center

Contrail boundaries and relative humidity with respect to ice (RHI) 1600 UTC, November 18, 2001 MODIS T 4 -T 5 Image Langley Research Center RHI from RUC-2 analysis, 225 mb

CONTRAIL OUTBREAK OVER GREAT LAKES, 9 OCT 2000 "Fly" aircraft through RUC fields, simulate formation & spreading Langley Research Center from Duda et al. 2003, JAS, accepted

Langley Research Center Garber et al. 2003, NASA

Comparison of contrail amount from satellite data and frequency of potential contrail conditions from RUC-2 data SEPTEMBER Langley Research Center OCTOBER

SUMMARY • Humidity fields critical for contrail formation, but correlations not always apparent because - extant cirrus may prevent detection of contrails from satellite contrail/cirrus conditions often equivalent - afternoon contrails may be less detectable because of overlap • 2001 coverage much less than 1993 -94 - 1993 -94 period one of moistest upper troposphere in 30 years (45. 5%) - 2001 one of driest at high altitudes in 30 years (39. 4%) - NOAA-16 may be less sensitive to contrails than NOAA-11 - NOAA-11 tendency for overestimation (~0. 5%) Langley Research Center

UPPER TROPOSPHERIC HUMIDITY OLD CONTRAILS ROADBLOCKS TO ACCURATE CONTRAIL PREDICTION THE AIR TRAFFIC SHUTDOWN CASE 2001 air traffic shutdown removed some impediments for contrail study - System cleared of commercial air traffic contrails by 0000 UTC, 12 September 2001 Langley Research Center

GOES-8 IR LOOP FOR SEPTEMBER 12, 2001, 1045 - 2345 UTC Langley Research Center

By studying the few contrails that occurred during the shutdown, we can tune a model that simulates contrails Langley Research Center

12 SEPTEMBER CONTRAIL ANALYSES • Use GOES images to track & compute spreading • Estimate heights using stereographic analysis - GOES-8 & AVHRR, MODIS, or GOES-10 - Flight levels between 10. 5 and 12. 5 km • Compute optical depth using RUC temperature at 11. 5 km (225 h. Pa) and adjacent clear-sky temperature Langley Research Center

3 -hour change in observed contrails NOAA-15 AVHRR 1 -km Infrared Image Terra MODIS 1 -km Infrared Image 1245 UTC, 12 September 2001 1545 UTC, 12 September 2001 H E Langley Research Center

DURING THIS EVENT MEAN CONTRAIL LIFETIME IS 6. 5 hr MEAN AREAL COVERAGE FOR EACH CONTRAIL IS 2270 km 2 MEAN OPTICAL DEPTH IS 0. 23 In the mean, over a 6. 5 hour period, 8 contrails covered 18, 000 km 2 => We need to relate the contrails to the humidity fields at altitude! Langley Research Center

Model humidity Langley Research Center

Comparison of RHI profiles from radiosondes & RUC-2 analyses 12 UTC, September 12, 2001 E, F, G A Langley Research Center B, C, D, H

NEED TO ARTIFICIALLY ENHANCE MOISTURE PROFILES OBSERVED & MODELED HUMIDITIES TOO LOW! RHI for adiabatic cirrus ~ 150% RHI for persistent contrails > 100% Nevertheless, there appears to be a relationship between the vertical structure of RHI and the lifetime, spreading, and optical depths of the observed contrails Langley Research Center

Comparison of persistent contrail occurrence and sonde RH, SLC, Utah Sassen, 1999, BAMS RHI correction for sonde profiles used here RHI corrections based on frost-point hygrometer data Miloshevich et al. 2001, JAS Langley Research Center

Radiosonde profiles of RHI in contrail areas after correcting for dry bias, 12 Z, September 12, 2001 Langley Research Center

Contrail simulation for September 12 over northeastern USA • Assume air traffic is equal to September 5 - use database from Garber et al. 2003 • Apply Appleman criteria, use RHI > 70% to define persistence • Assume RHI for a flight track = that of nearest RUC level • Compute spreading, assuming fall speed of 3 cm/s - max width = 12 km, shear determines spread rate (mean = 6 km/h, same as observed for military contrails) - no new nucleation, optical mass = OD*width - OD = f(t), peaks at 2 hours • Advect old & add new contrails • Delete trails if RHI < 70% or older than 6 hr Langley Research Center

Hourly simulation of contrails over northeastern USA September 12 2001, assuming air traffic for September 5 Langley Research Center

ANOTHER VIEW Langley Research Center

Estimate of linear contrail coverage over northeastern US for September 12, 2001 assuming air traffic for Sept. 5 Coverage could have been at least 200, 000 km 2 if traffic were normal Daily global mean is 200, 000 - 400, 000 km 2! Langley Research Center

ALTERNATIVES FOR MITIGATION • Predict locations & altitudes where contrails most likely - provide alternate routing (height change) - our proposal • Fly lower all the time - German study 2003 • Use liquid hydrogen fuels - German study 2003 Langley Research Center

Langley Research Center

Langley Research Center

Langley Research Center

Langley Research Center

CONCLUDING REMARKS • Understanding has increased rapidly but remaining issues - accuracy of contrail coverage - where does contrail impact occur how much is local? spread around? - contrail optical depths 0. 1 - 0. 3? geographic dependence? - indirect effects on cirrus clouds does AC exhaust increase opt depth, decrease De? do AC exhaust aerosols reduce formation threshold leading to more cirrus? - is there an effect on the moisture budget? - how accurately can we model cirrus aviaticus? need good natural cirrus model (satellite comparisons) - what are best Langley Research Centermitigation options?

FUTURE RESEARCH AT LANGLEY • Continue climatology development - examine relationship of natural cirrus to environment • Compute radiative forcing for simulations • Apply more sophisticated contrail model, account for natural Ci - match flights to vertical details of RHI - improve precip, spreading, dissipation • Improve threshold for contrail persistence from models - find a new data source (RUC changed April 19) • Determine source of differences in lifetimes and spreading • Push for improved UT RH • Provide realistic real time predictions of contrails in clear air Langley Research Center

REFERENCES & MISCELLANEOUS • All references can be found on our main web page http: //www-pm. larc. nasa. gov/ click on "SASS", then on "Related References" • Other imagery and examples are available on the same main web page, click on "PATHFINDER", "SUCCESS" or "SASS" • A near-real-time contrail predictor is available on the main web page, click on "Contrail Forecast" (not so good since RUC change) Langley Research Center