![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081105140114im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Detection of a Recovery in Stratospheric and Total Ozone
EPA Grant Number: R829402C001Subproject: this is subproject number 001 , established and managed by the Center Director under grant R829402
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: Center for Integrating Statistical and Environmental Science
Center Director: Stein, Michael
Title: Detection of a Recovery in Stratospheric and Total Ozone
Investigators: Tiao, George , Fioletov, Vitali , Flynn, Lawrence , Guillas, Serge , Hayhoe, Katharine , Kerr, James , Meng, Xiao-Li , Miller, Alvin , Petropavlovskikh, Irina , Reinsel, Gregory , Weatherhead, Elizabeth , Wuebbles, Donald J. , Yang, Shi-Keng
Institution: Environment Canada , Harvard University , National Oceanic and Atmospheric Administration , University of Chicago , University of Colorado , University of Illinois at Urbana-Champaign , University of Wisconsin - Madison
EPA Project Officer: Smith, Bernice
Project Period: March 12, 2002 through March 11, 2007
RFA: Environmental Statistics Center (2001)
Research Category: Ecological Indicators/Assessment/Restoration , Environmental Statistics
Description:
Objective:This project had a more applied focus than some of the other projects supported by CISES and mainly addressed the appropriate application of statistical methods to environmental problems, leading largely to publications in major science journals rather than statistical journals. This research aimed at addressing specific substantive issues and changing statistical practices in the environmental sciences. In a larger sense, the group sought to resolve a fundamental problem within the scientific community in that there has been too little science in most environmental statistics and too little statistics in most environmental science. The net result is that the relevant environmental community now utilizes these statistical elements as standard whereas they were very much unaware of them only a few years ago. In keeping with its applied emphasis, the personnel associated with this project represented a consortium of statisticians, physicists and meteorologists who brought their specialties to a synergistic program to resolve the statistical issues raised by the project’s title.
The importance of this work to EPA was acknowledged in 2005 when the Science Team on this project was awarded an EPA Stratospheric Ozone Protection Award. In presenting us with the award, the EPA stated: “The Ozone Science Tiger Team, a consortium of U.S and Canadian universities and government laboratories, has played a unique and major role in documenting human impacts on global tropospheric and stratospheric ozone. The Tiger Team of renowned statisticians and atmospheric scientists launched an innovative, cross-disciplinary approach to studying ozone trends and has engaged the public and guided policy to protect stratospheric ozone. Through its ongoing research, numerous publications, and active participation at national and international forums, the Tiger Team continues its vital role of informing ongoing scientific and policy discussions on ozone layer recovery and impacts from climate change.”
Summary/Accomplishments (Outputs/Outcomes):1.1 Overall summary
There has been a major change in the ozone trend dating to roughly 1996. A reasonable modeling scenario is to utilize a trend since 1979 with a trend-change-point in 1996 and to remove 2 years of data post Mt. Pinatubo eruption. We found that including a term for the recognized source of atmospheric variability known as the “Arctic Oscillation” within the statistical model yielded a statistically significant effect and contributed about 25% of the variance. Thus, inclusion of stratosphere-troposphere linkages is an important aspect that must be included within the physical models. Finally, the statistical mechanism offered provides a continuing framework for examination of the anticipated ozone recovery.
In addition, statistical approaches were developed to improve capabilities for analyzing computer models of atmospheric chemistry and physics. Guillas et al. (2006) introduced a hybrid statistical chemical-transport model for total column ozone prediction, based on the University of Illinois at Urbana-Champaign 2-D (UIUC 2-D) chemical-transport model of the global atmosphere. The 1996-2003 validation sample confirmed that the combined approach yields better predictions than the direct UIUC 2-D outputs.
Lastly, there is a larger question as to how far in the future it will take to determine that any variable (e.g. surface temperature) has changed with satisfactory statistical significance. Toward this, George Tiao and his students developed a generic statistical program that allows researchers to answer the above. A paper outlining the details of this program is in preparation and the program will be made available via open source software.
1.2 Ozone recovery in the lower stratosphere
One major question that arises with the implementation of the Montreal Protocol in 1985 and its subsequent Conventions is our ability to determine that an ozone “recovery” is in process. Toward this we utilized a statistical model suggested by Reinsel et al. (2002) that utilizes the idea of a trend and a trend-change at a specific time and applied it to both global total column ozone derived from the NOAA Solar Backscattered Ultraviolet Ozone Sensor (SBUV) (Miller et al., 2001; Fioletov et al., 2002) and 12 balloon ozonesonde stations in the mid-latitudes of the northern hemisphere. While the satellite data provide a global perspective of the total column ozone, the lower stratosphere is of particular significance as this is where the ozone concentration is a maximum and also where heterogeneous ozone losses have been noted. This statistical methodology offers independence from any particular forecast model, but suffers from the ambiguities of having to select a specific time for the ozone trend to change and the fact that the Mt. Pinatubo volcanic aerosols in 1991 impacted the ozone amount.
Within Miller et al. (2006) we analyzed the ozonesonde station data utilizing the above model, but examined the statistical stability of the computed results by allowing the point of inflection to change from 1995 through 2000 and also excluded varying amounts of data from the post-Pinatubo period. After examining the utilization of the trend and trend-change methodology on the northern hemisphere mid-latitude ozonesonde stations, along with the sensitivity of the assumptions, the results can be summarized for the lower stratosphere as:
- The impacts of deleting up to 4 years of data and changing the inflection point from 1995-2000 are non-trivial, but the overall results are consistent.
- There has been a major change in the ozone trend within the time-frame of 1996 and the ozone in the lower stratosphere is increasing from that approximate time.
- A reasonable scenario is to utilize a change-point in 1996 and remove 2 years of data post Mt. Pinatubo eruption.
- Including a term for the known atmospheric variability known as the “Arctic Oscillation” (Miller et al., 2003; Wieliki et al., 2002) within the statistical model demonstrates that it is statistically significant and contributes about 25% of the variance. Thus, inclusion of stratosphere-troposphere linkages is an important aspect that must be included within the physical models. The role of the Arctic Oscillation in the atmosphere impacting Project 2 will be demonstrated below.
- It provided a continuing framework for examination of the anticipated ozone recovery. Several other statistical mechanisms are based on specific computer simulations. The Trend and Trend- Change model is independent of computer simulations and can be utilized into the future to test observed changes versus those anticipated.
- It showed a very clear statistical relationship between the tropospheric forcing in the form of the surface Arctic Oscillation and the stratospheric ozone. This requires follow-up in that this concept can be important in either the influence of the troposphere on stratospheric changes or the vice-versa in that the stratosphere influences the troposphere on time scales important for forecasting, e.g. out to two weeks (e.g. Zhou et al., 2001).
- The results suggested a pattern of solar response of ozone as a function of altitude that does not agree with current hypotheses. The influence of the 11-year solar cycle on the earth’s surface has been the subject of considerable speculation with significant impact on long-term outlooks if proven. This must be examined further to substantiate the results as these data could serve as a true independent test of several hypotheses of solar-surface relationships.
This paper’s importance lies in several areas:
1.3 Ozone recovery in total column ozone
The above statistical analysis techniques were applied to the total column ozone with results that support the above conclusions. Not only does this substantiate the general statistical conclusions, but also serves as a model for generic analysis of “ozone recovery” in the future.
Reinsel et al. (2005) was a key paper in such analyses (it was also selected by Discovery Magazine as one of the top science results of 2005). Statistical trend analyses were performed for monthly zonal average total ozone data from both TOMS and SBUV satellite sources and groundbased instruments over the period 1978-2002 for detection of a “turnaround” in the previous downward trend behavior and hence evidence for the beginning of an ozone recovery. Since other climatic and geophysical changes can impact ozone behavior and can influence the detection of turnaround and recovery, we also focused on accounting for ozone variations that may be ascribed to various physical and chemical influences. Thus we included in the statistical trend modeling and analysis the effects of various dynamical and circulation variations in the atmosphere, including those associated with the quasibiennial oscillation (QBO), Arctic Oscillation (AO) and Antarctic Oscillation (AAO), and Eliassen-Palm (EP) flux influences, as well as influences of solar cycle. A notable result of the analysis was that for latitude zones of 40± and above in both hemispheres, large positive and significant estimates of a change in trend (since 1996) were obtained (on the order of 1.5 to 3 DU per year). The dynamic index series, AO/AAO and EP flux, were found to have a substantial influence on total ozone for these higher latitudes, and significant influences of lesser magnitude were also found for lower latitudes. The feature of positive significant change in trend in total ozone over recent years, however, was obtained both without and with the dynamical index terms included in the statistical models.
1.4 A generalized statistical analysis program
In addition to the basic question as to whether or not there has been a statistically significant change in ozone, there is a larger question as to how far in the future it will take to determine that any variable (e.g. surface temperature) has changed with satisfactory statistical significance. Toward this, George Tiao and his students developed a generic statistical program that allows researchers to answer the above. A paper outlining the details of this program is in preparation and the program will be made available via open source software. This program grows out of previous studies done for the CISES program by Tiao and by Elizabeth Weatherhead.
1.5 Explaining past trends in stratospheric ozone
Guillas et al. (2004) combined observations with atmospheric model results to make a unique analysis of past trends in ozone and what these trends imply about future recovery of stratospheric ozone. Since the implementation of the international controls on ozone depleting chemicals, an important focus in studies of stratospheric ozone has been on the detection of a turnaround in the downward trend. The usual statistical assumption of a piecewise linear representation of the trend does not account for the chemistry involved. Because the actual change is not expected to be piecewise linear, we modeled the trend using results from a University of Illinois at Urbana- Champaign 2-D (UIUC 2-D) chemical-transport model of the global atmosphere. As part of the analysis, ozone observations were considered in the spectral domain using a cohesive data set from the SBUV-SBUV/2 satellite system at northern midlatitudes. In this study, we found that the new model is better at capturing the long range correlation of the data than assuming a linear trend. We also compared several statistical trend models, based either on a regression on a linear trend, or on the Effective Equivalent Stratospheric Chlorine (EESC). Including a constant halocarbon emissions run of the UIUC 2-D model in the regression, the controlled EESC approach showed the best fit. This demonstrated the ability of the model to simulate the natural variation of total column ozone. The smallest future data length necessary to detect a recovery with a certain probability was obtained in this latter case.
For the studies here to evaluate how well the UIUC 2-D Chemical-Transport Model can represent observed trends in stratospheric ozone, we included as inputs to the model:
- The solar mid- and near-ultraviolet radiation (200 to 400nm) data from 1970 to 2002 (as updated by J. Lean, private communication, 2003, based on earlier analyses in Lean et al., 1997);
- The surface mixing ratios from WMO (2003) for all relevant source gases (e.g., halocarbons, methane, nitrous oxide, carbon monoxide) but updated using recent data from Climate Monitoring & Diagnostics Laboratory (CMDL, using data on http://www.cmdl.noaa.gov/
); - Stratospheric aerosol surface area densities derived from extinction measurements by Stratospheric Aerosol and Gas Experiment (SAGE) II from 1981 to 2002 (as updated by L. Thomason, private communication, 2003, based on earlier analyses in Thomason et al., 1997);
- Tropospheric and stratospheric temperature data from 1980 to 2002 based on a combination of analyses for 1000 hPa to10 hPa based on the National Center for Environmental Prediction / National Center for Atmospheric Research (http://dss.ucar.edu/pub/reanalysis/ ) reanalysis, and the data from 10 hPa to 0.5hPa based on NOAA Climate Predict Center (CPC; Alvin J. Miller, private communication, 2003) analyses.
1.6 Diagnostics for atmospheric models
Statistical approaches were developed (e.g., Guillas et al., 2006) to improve capabilities for analyzing numerical models of atmospheric chemistry and physics. Guillas et al. (2006) introduced a hybrid statistical chemical-transport model for total column ozone prediction, based on the University of Illinois at Urbana-Champaign 2-D (UIUC 2-D) chemical-transport model of the global atmosphere. We proposed a general diagnostic procedure for the model outputs in total ozone over the latitudes ranging from 60±S to 60±N to see if the model captures some typical patterns in the data. The method proceeded in two steps to avoid possible collinearity issues. First, we regressed the measurements given by a cohesive data set from the SBUV(/2) satellite system on the model outputs with an autoregressive noise component. Second, we regressed the residuals of this first regression on the solar flux, the annual cycle, the Antarctic or Arctic Oscillation, and the Quasi Biennial Oscillation. If the coefficients from this second regression were statistically significant, this meant that the model did not simulate properly the pattern associated with these factors. Systematic anomalies of the model were identified using data from 1979 to 1995, and statistically corrected afterwards. The 1996-2003 validation sample confirmed that the combined approach yields better predictions than the direct UIUC 2-D outputs.
This very unique way of testing and evaluating physics and chemistry based models of the atmosphere and other components of the Earth system has now been applied to an air quality model at Georgia Tech University. Serge Guillas and Don Wuebbles continue to look at new ways to apply statistical techniques to the evaluation of large complex numerical models.
There is an increasing interest in studying time-varying quantiles, particularly for environmental processes. For instance, high pollution levels may cause severe respiratory problems, and large precipitation amounts can damage the environment, and have negative impacts on the society. Draghicescu, Guillas and Wu (2008) addressed the problem of quantile curve estimation for a wide class of non-stationary and/or non-Gaussian processes. We discussed several nonparametric quantile curve estimates, gave asymptotic results, and proposed a data-driven procedure for the selection of smoothing parameters. This methodology provides a statistically reliable and computationally efficient graphical tool that can be used for the exploration and visualization of the behavior of time-varying quantiles for non-stationary time series. A Monte Carlo simulation study and two applications to ozone time series illustrated our method.
Conclusions:Contributions to understanding of environmental problems
Trend estimation is fundamental to assessing the state of the environment and the effectiveness of policies intended to affect the environment. Trend estimates without appropriate estimates of uncertainty are not particularly helpful, so statistical methods for producing valid uncertainty estimates are essential. This project built on a long-term collaboration of academic and government scientists known as the Tiger Team that examined trends in stratospheric ozone levels. Hallmarks of this group’s work include careful treatment of data, the use of a wide range of information sources for explaining natural variability in ozone levels and the development of appropriate but widely usable methods for accounting for dependencies across time in the ozone levels.
One specific issue addressed in this project was the development of statistical methods for detecting and estimating changes in ozone trends due, in particular, to the implementation of the Montreal Protocol in 1985. We found evidence of a change in ozone trend starting around 1996, with the previous trend of decreasing ozone levels now replaced by an increasing trend.
Another specific issue was the inclusion of various indices of atmospheric circulation in models for ozone trends. To the extent that variations in ozone levels can be explained by these indices, it is possible to get estimates of ozone trends with less uncertainty. The quasibiennial oscillation (QBO) is probably the most famous of these indices, but we found that some less familiar indices such as the Arctic Oscillation have a substantial influence on ozone levels at higher latitudes. We also explored the use of output of computer models of the upper atmosphere to explain variability in ozone levels and found this approach could shorten the time needed to detect a recovery in ozone levels considerably.
By using carefully chosen but not overly complicated methods, our methods have already gained wide use in ozone trend estimation. Our methods, however, are not specific to ozone and can be applied to a broad range of problems in the estimation and detection of trends in environmental indicators.
References:
Fioletov, V.E., G.E. Bodeker, A.J. Miller, R.D. McPeters, and R. Stolarski, 2002: Global and zonal total ozone variations estimated from ground-based and satellite measurements:1964-2000, J. Geophys. Res., 107, 4647, doi:10.1029/2001JD001350.
Lean, J. L., G. J. Rottman, H. L. Kyle, T. N. Woods, J. R. Hickey, and L. C. Puga, 1997: Detection and parameterization of variations in solar mid- and near-ultraviolet radiation (200-400 nm), J. Geophys. Res. 102, 29939-29956.
Miller, A.J., S. Zhou and S.K. Yang, 2003: Relationship of the Arctic and Antarctic Oscillation to the outgoing longwave radiation, J. Climate, 16, 1583-1592.
Miller, A.J., R. M. Nagatani, L. E. Flynn, S. Kondragunta, E. Beach, R. Stolarski, R. McPeters, P. K. Bhartia , M. DeLand, C.H. Jackman , D.J. Wuebbles, K.O. Patten and R.P. Cebula, 2002: A Cohesive Total Ozone Data from the SBUV(/2) Satellite System, J. Geophys. Res., 107, doi:10.1029/2001JD000853.
Reinsel, G. R., A. J. Miller, L.E. Flynn, R. M. Nagatani, G. C. Tiao, E. C.Weatherhead, and D. J. Wuebbles 2005: Trend analysis of total ozone data for turnaround and dynamical contributions. J. Geophys. Res., 110, doi: 10.1029/2004JD004662
Thomason, L. W., L. R. Poole, and T. R. Deshler, 1997: A global climatology of stratospheric aerosol surface area density as deduced from SAGE II:1984-1994, J. Geophys. Res. 102, 8967-8976.
Wielicki, B.A. et al., 2002: Evidence for large decadal variability in the tropical mean radiative energy budget, Science, 294, 841-844. Greece, May 31-June 8, 2004.
Zhou, S., A. J. Miller, J. Wang and J. K. Angell, 2001: Trends of NAO and AO and their associations with stratospheric processes, Geophys. Res. Lett., 28, 4107-4110.
Journal Articles on this Report: 4 Displayed | Download in RIS Format
| Other subproject views: | All 26 publications | 4 publications in selected types | All 4 journal articles |
| Other center views: | All 102 publications | 59 publications in selected types | All 37 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Guillas S, Stein ML, Wuebbles DJ, Xia J. Using chemistry transport modeling in statistical analysis of stratospheric ozone trends from observations. Journal of Geophysical Research 2004;109(D22303), doi:10.1029/2004JD005049. |
R829402C001 (2004) R829402C001 (Final) R829402C002 (2004) |
|
|
|
Guillas S, Tiao GC, Wuebbles DJ, Zubrow A. Statistical diagnostic and correction of a chemistry-transport model for the prediction of total column ozone. Atmospheric Chemistry and Physics 2006;6(2):525-537. |
R829402C001 (2004) R829402C001 (Final) |
|
|
|
Miller AJ, Cai A, Tiao G, Wuebbles DJ, Flynn LE, Yang S-K, Weatherhead EC, Fioletov V, Petropavlovskikh I, Meng X-L, Guillas S, Nagatani RM, Reinsel GC. Examination of ozonesonde data for trends and trend changes incorporating solar and Arctic oscillation signals. Journal of Geophysical Research 2006;111(D13305), doi:10.1029/2005JD006684. |
R829402C001 (2004) R829402C001 (2006) R829402C001 (Final) |
|
|
|
Reinsel GC, Miller AJ, Weatherhead EC, Flynn LE, Nagatani RM, Tiao GC, Wuebbles DJ. Trend analysis of total ozone data for turnaround and dynamical contributions. Journal of Geophysical Research 2005;110(D16306), doi:10.1029/2004JD004662. |
R829402C001 (2004) R829402C001 (Final) |
|
, Ecosystem Protection/Environmental Exposure & Risk, Economic, Social, & Behavioral Science Research Program, Air, Geographic Area, Scientific Discipline, Health, RFA, PHYSICAL ASPECTS, Ecosystem/Assessment/Indicators, Engineering, Chemistry, & Physics, Risk Assessments, decision-making, Environmental Statistics, Great Lakes, Applied Math & Statistics, Health Risk Assessment, Physical Processes, Ecological Risk Assessment, Environmental Engineering, Ecological Effects - Human Health, EPA Region, particulate matter, Ecological Effects - Environmental Exposure & Risk, Ecosystem Protection, Monitoring/Modeling, Environmental Monitoring, risk assessment, trend monitoring, ozone , chemical transport models, particulate, risk management, stochastic models, statistical methodology, air quality, computer models, ecological risk, ecosystem health, environmental indicators, ozone, chemical transport, health risk analysis, human health risk, monitoring, policy making, statistical models, particulates, regulations, statistical methods, watersheds, Region 5, air pollution, stratospheric ozone, data models, exposure, water, chemical transport modeling, ecological models, ecological effects, ecological health, human exposure
Progress and Final Reports:
2002 Progress Report
2004 Progress Report
2006 Progress Report
Original Abstract
Main Center Abstract and Reports:
R829402 Center for Integrating Statistical and Environmental Science
Subprojects under this Center:
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R829402C001 Detection of a Recovery in Stratospheric and Total Ozone
R829402C002 Integrating Numerical Models and Monitoring Data
R829402C003 Air Quality and Reported Asthma Incidence in Illinois
R829402C004 Quasi-Experimental Evidence on How Airborne Particulates Affect Human Health
R829402C005 Model Choice Stochasticity, and Ecological Complexity
R829402C006 Statistical Approaches to Detection and Downscaling of Climate Variability and Change