Current Debate on Stratospheric Temperature Trends from SSU

Current Debate on Stratospheric Temperature Trends from SSU Cheng-Zhi Zou NOAA/NESDIS/Center For Satellite Applications and Research With Help from Haifeng Qian, Wenhui Wang, Likun Wang 2013 GSICS Annual Meeting Williamsburg, Virginia, 4 -8 March 2013

Introduction · Ø Ø Stratospheric temperature trend is an important indicator of anthropogenic global warming Stratospheric cooling: Ozone depletion Increasing carbon dioxide and other greenhouse gases · Radiosonde observations can’t reach to mid-upper stratospheres · Lidar observations are sparse · Rely on satellite observations

The SSU Instrument · One of the NOAA TOVS instruments (MSU, HIRS, SSU) from 1978 -2007 · Infrared radiometer use pressure modulation technique to measure atmospheric radiation from CO 2 15 - m v 2 band · An interference filter allows only 15 - m band to pass through · A cell of CO 2 gas is placed in the instrument’s optical path with its pressure changed in a cyclic manner

Channels · Weighting function determined by the pressure values · Three different pressures to give three different weighting function Channel Number Central Wave Cell pressure weighting function No. (cm-1) (pre-launch peak (wavelength) specific) 1 668 (15 m) 100 (h. Pa) 15 mb (29 km) 2 668 (15 m) 35 (h. Pa) 5 mb (37 km) 3 668 (15 m) 10 (h. Pa) 1. 5 mb (45 km) P(peak)~P(cell)/[CO 2]1/2,

Scan and Calibration Cycle

SSU Satellites Only 7 among the 9 NOAA TOVS satellites were equipped with SSU 6

Brightness Temperature Anomalies—From NOAA Operational Calibration • 5 -day and global averaged Tb anomaly time series • Include all 8 pixels per scanline • Global coverage • Cloud effect minimal ; include most observations • Global inter-satellite differences between NOAA-7 and NOAA-8 are as large as 4 K El Chicon Mt. Pinatubo 7

SSU Data Issues Ø Gas leaking problem in the CO 2 cell pressure change Ø atmospheric CO 2 variations Ø limb-effect Ø diurnal drift effect semi-diurnal tides Ø inter-satellite biases Ø No instruments on NOAA-10 and NOAA-12 Ø No overlap between NOAA-9 and NOAA-11

Cell Pressure Time Series from Gas Leak 9

Effect of Cell Pressure Decreasing CO 2 cell pressure decreasing -> weighting function peaks higher -> because of increasing lapse rate, measured Tb increasing -> warm bias SSU cell pressure Measured BT 10 Weighting function

SSU CDR Development Flow Chart Original SSU BTs Interpolated profiles along each SSU pixels SSU BTs with fixed cell pressures SSU BTs with removing CO 2 increasing effects SSU BTs with limb adjustments SSU BTs with diurnal corrections SSU simulations with real and fixed cell pressures SSU simulations with fixed and varying CO 2 amount SSU simulations at nadir and offnadir Diurnal correction database Well-merged SSU gridded BTs SSU TCDR

Reanalysis Uncertainties Don’t have an Impact on Correction § Bias corrections are sensitive to the layer temperature differences, not the temperature profile itself § The spurious temperature jumps near 1996 and 2000 are apparently due to model errors § These jumps in MERRA temperatures did not show up in the simulated correction time series (bottom panel) § Demonstrating reanalysis uncertainties may not be critical to the accuracies of the corrections No big jumps between satellites and streams in MERRA Top: CRTM simulated channel 3 global mean time series for NOAA 11 (red) and NOAA 14 (blue) with varying CO 2 cell pressure (color) and fixed CO 2 cell pressure (grey). Bottom: Their differences (color minus grey)

Effect of Correction and Merging After instrument CO 2 cell + atmospheric CO 2 correction, the original upward trend became flat for ch 2 and ch 3 NOAA -7 biases were reduced after CO 2 cell correction After instrument CO 2 cell + atmospheric CO 2 correction, the original downward trend became even more negative for ch 1

The Trend Debate Plot from Thompson et al. 2012 in Nature 14

Debate Example #1 Ø UKMO did not correct Cell pressure effect for Channel 1 Ø But NOAA did NOAA-9, NOAA-11 15 NOAA-9 NOAA-11

Debate Example #2 SSU channels 2, 3, and MSU channel 4 all flat from 1985 -1992 Only SSU channel 1 shows downward trend-- inconsistency 16

Debate Example #3 Disconnection problem for channel 3 Is the large drop (0. 3 K in 6 months) real or bad observations at the early stage of NOAA-14 17

Debate Example #4 Channel 3 global mean trend is similar between NOAA and UKMO 18 But zonal mean trends have very different patterns

Ongoing Activity for improvements p Collect more information from other scientists: Roger Saunders, John Nash, Jim Miller, Mike Chalfant, Tony Reale, Laurie Rokke, Shinya Kobayashi, Mitch Goldberg…. p p Modeling climate Community is joining the debate: Adrian Simmons, Dick Dee … Level-1 c calibration– NOAA operational calibration was used before. We are trying to develop new calibration schemes to understand if it can make differences; Especially for inconsistencies between channels (Reprocessing level-1 c is time consuming and need more support) p CRTM – not sure if inconsistency between channels are related to CRTM, but check with CRTM teams (Mark Liu, Yong Chen) to see if there are potential problems in CRTM p Checking diurnal drift corrections– maybe related to channel 3 zonal mean trend pattern?

References p Wang, L. , C. -Z. Zou, and H. Qian (2012), Construction of stratospheric temperature data records from Stratospheric Sounding Units. J. Climate, Vol 25, 2931 -2946 p Thompson, D. W. J. , D. J. Seidel, W. J. Randel, C. Z. Zou, A. H. Butler, C. Mears, A. Osso, C. Long, R. Lin, (2012): The mystery of recent stratospheric temperature trends. Nature, 491, 692 -697. doi: 10. 1038/nature 11579 p Zou, C. -Z. , et al. (2013) On the differences of SSU datasets between the NOAA and UK Met Office versions, In preparation

Thank You! 21