AIM - The Mission

Images from the CIPS instrument have been used to generate the first maps of daily cloud brightness at high spatial resolution for the northern and southern 2007 seasons, to determine mesospheric ice particle properties and ice mass, and to unveil new cloud features. An example of a daily cloud map in the heart of the cloud season is shown in Figure 2.3.1-1. The pixel size is 5 x 5 km, making this the best spatially resolved cloud image ever taken from space. Similar maps are generated each day of the cloud season and show the time evolution of cloud presence in great detail. The hole over the pole is due to lack of data.

In addition to the clear gravity wave features in the CIPS scenes, cloud features such as 'ice rings' and spatially small but bright clouds have prompted a fresh look into the dynamics controlling the summer mesosphere. The ice ring features (see Figure 2.2.2-1) can be caused by gravity waves, but may also be due to turbulent mixing from vertically displaced air. Ice rings appear in circular or oval shapes with diameters that vary from ten to several hundred kilometers or greater. They have been photographed from the ground, but never explained. Another indication of dynamical control of the summer mesosphere is seen in the relatively frequent appearance of spatially small (~20 - 30 km diameter) but bright clouds. The cloud brightness may approach four times the Rayleigh background at small scattering angles. These bright clouds have not been reported previously and may be a manifestation of an intense small-scale upwelling. These represent still another unsuspected aspect of the complicated dynamics that controls transport and temperature in the heart of the cloud season.

SOFIE is producing unprecedented measurements of key parameters from the lower stratosphere up to 95 km and higher, depending on product. Temperature, H2O, CH4, and O3 are included in the January 2008 release, with CO2 and NO to be added in the next release. The time versus height cross sections of SOFIE H2O in Figure 2.2.1-1 illustrates the development of water vapor versus altitude in the PMC environment during the 2007 northern cloud season. Water vapor, an essential gas for PMC formation, is measured at a vertical resolution of 1.6 km and precision of less than 21 ppbv. The seasonal evolution of water vapor is characterized by a steady increase in mixing ratio, which is caused by vertical transport due to upwelling, and also by sublimation of PMC particles below ~83 km. The abrupt decrease in H2O above PMC altitudes is consistent with a loss of water vapor due to ice growth, and also with destruction of H2O by increasing photolysis at higher altitudes. An example of the quality of the underlying profiles is demonstrated in Figure 2.3.2-1, which shows a typical set of retrieved water vapor profiles, in this case for August 12, 2007. The various profiles indicate where PMCs have formed as evident by reduced water vapor near 85 km, and where PMCs have sublimated resulting in enhanced water vapor just above 80 km. The mesospheric ozone minimum is observed near 79 km altitude, with a peak observed near 90 km (Figure 2.2.2-1). O3 enhancement is observed at altitudes from roughly 80 to 95 km during times that ice is present. For example, at 92 km the mixing ratio increases from 0.8 ppmv on 22 May (30 days before solstice) to ~1.1 ppmv by 11 July, returning to ~0.8 ppmv by 15 August. Siskind et al. [2007] discuss the increase in O3 above PMCs and determined that the change was consistent with dehydration above the ice layer resulting in lowered HOx. These observations of highly transient features, combined with the other SOFIE products are giving a first-time detailed look at the physics of PMCs.

The multi-wavelength SOFIE observations are shedding new light on PMCs with unprecedented sensitivity and signal-to-noise. While previous instruments suggested that ice is present about half the time near 70?N, SOFIE now shows that ice is almost always present at these latitudes during the PMC season. This is demonstrated in Figure 2.3.3-1a, where a time series of SOFIE ice occurrence frequency is compared to SME and HALOE results. Note that downgrading SOFIE sensitivity to be consistent with HALOE and SME yields good agreement between SOFIE and the SME and HALOE frequencies, confirming that the lower ice frequencies recorded by the older satellites were a result of reduced sensitivity. SOFIE has also demonstrated that mesospheric ice exists as a continuous layer from ~80 km to altitude to the mesopause (~87 km) and often higher, as shown in Figure 2.3.3-1b. This view is supported by model predictions [Rapp and Thomas, 2006], but is in contrast with previous measurements that were not sensitive enough to detect the most tenuous ice populations.

These new discoveries were possible because SOFIE is 15 - 200 times more sensitive to PMC particles than previous instruments, including lidars. SOFIE measures PMC extinction at 11 wavelengths from 0.330 to 5.006 ?m, with 1.5 km vertical resolution and a precision of better than 10-7 km-1 in extinction. In addition, SOFIE contains channels located at the peak in the ice absorption spectrum (~3 ?m wavelength) where PMC extinction is ~100 times greater than for surrounding NIR and IR wavelengths. The unique combination of measurement wavelengths allows SOFIE observations to be used to determine vertical profiles of PMC ice mass density, particle shape, and Gaussian size distributions.

Time series of SOFIE ice layer altitudes, occurrence frequency, mass density, particle shape (axial ratio of a spheroid), effective radius, and water vapor are also shown in Figure 2.3.3-1 for 69?N average latitude. Ice layers with greater vertical extent are found to have higher mass densities, which supports current growth /sedimentation theories which suggest that thicker saturated regions that are thicker in vertical extent will provide increased growth time for falling particles. SOFIE offers the first observations of PMC particle shape versus altitude. SOFIE observations indicate that PMC particles are non-spherical at all altitudes, with axial ratios of about 2.2 (i.e., oblate spheroids) at the PMC extinction peak (Zmax). Examining particle shape as a function of altitude reveals axial ratios in excess of 3 at altitudes above and below the extinction peak, and a strong anti-correlation to ice mass density (Figure 2.3.3-2). These new results suggest that changes should be made to the current treatment of microphysical processes including particle sedimentation and possibly nucleation and growth. SOFIE indicates PMC particle effective radii at Zmax increasing from ~20 nm in the early season to over 40 nm by early August. A correlated and steady increase is observed in water vapor measured at Zmax (Figure 2.3.3-1), in conjunction with decreasing particle concentrations (not shown). A robust relationship between particle size and water vapor is found to hold when examining variability over time at one altitude, and when examining the altitude dependence of H2O and particle size. These findings point to a fundamental connection between ice particle characteristics and water vapor. Increased water vapor near 83 km is expected due to upwelling and also from ice sublimation [Hervig et al., 2003], and it now appears that that increased H2O is in turn modifying the ice particles characteristics. SOFIE results have been compared to concurrent PMC measurements from the ALOMAR lidar (69?N) during 2007 [Baumgarten et al., 2008], which were used to determine various PMC properties including Gaussian size distributions. Example results are shown in Figure 2.3.3-3, where histograms of the Gaussian size distribution parameters determined at the extinction peak altitude from SOFIE and the lidar are compared. The agreement is excellent for particle concentration, median radius, and distribution width. The notable difference is that SOFIE indicates the presence of smaller particles at higher concentrations than the lidar, a difference that is consistent with SOFIE having greater sensitivity to smaller particles. SOFIE is also in good agreement with the range of values indicated by model results and previous lidar measurements, as shown in Figure 2.3.3-3.

(on-line bibliography)
The AIM data processing and validation has been moving at a faster pace than anticipated based on past experience with large satellite data sets. Even though launch occurred less than 10 months ago, Nine AIM papers were presented at the international conference on Layered Phenomena in the Mesopause Region in Fairbanks Alaska in August, 2007 and 15 were presented at the 2007 Fall AGU meeting. The fact that so many papers could be presented so soon after launch attests to the high quality of the AIM results. In addition to these presentations, 14 papers have been submitted for publication in the Journal of Atmospheric and Terrestrial Physics. A full AIM post-launch bibliography and submitted versions of the journal articles are available on the AIM web site at “aim.hamptonu.edu”.

The AIM mission was only recently launched, so there are no citations of papers published. However, the mission has received extensive media attention world wide both before and after launch. A press conference was held at the 2007 Fall AGU meeting that was attended by approximately 15 reporters including major news organizations such as the AP, Reuters and the BBC. The conference went for about an hour and had to be terminated because of room use. The mission has been publicized broadly throughout the world. A few of the many news organizations that have published AIM stories either online, in print, or both include: Space News, Chemical & Engineering News, The BBC World News-Online, Cosmos Magazine in Australia, Forbes-Online, MSNBC-Online, The Examiner-Online (Philadelphia and Minneapolis), The Baltimore Examiner, CBSNews.com, Los Angeles Times- Online, Houston Chronicle- Online, Washington Post- Online, the New York Times, USA Today – Online, Northwest Florida Daily News, Discovery Channel TV in Canada, KFMB-TV (CBS affiliate- San Diego, CA), WFED-AM (Washington D.C.), WTOP-FM (Washington D.C.), WVEC-TV (Norfolk, Virginia), WHEC-TV (NBC affiliate -Rochester, NY), KAAL-TV (ABC affiliate – Rochester-MasonCity-Austin, IA –MN), and KSL-TV (NBC affiliate Salt Lake City, UT). The AIM mission was a focus at the Fairbanks, Alaska meeting in August, 2007 on Layered Phenomena in the Mesopause Region. Also, even though it was a late breaking session for the 2007 Fall AGU meeting, the noctilucent cloud session attracted over 40 papers of which 15 were addressing AIM.

Achieving this objective as well as all other AIM objectives requires, first and foremost, that the AIM data be validated. Substantial progress has been made toward this goal in the eight months since data were first acquired. For example, in Figure 2.6.1-1 CIPS and SBUV PMC brightness measurements are compared. For this analysis, CIPS data were binned to match the SBUV field of view, and a cloud detection algorithm based on the SBUV algorithm [DeLand et al., 2003] was applied. The agreement is excellent, with average differences of only 2%; we thus have confidence that the CIPS data are valid for scientific investigations [Benze et al., 2008].

Detailed microphysical models suggest that after nucleation, PMC cloud particles will grow large enough to fall into a region of warmer temperatures where they sublimate. The resultant evaporated H2O can then be re-lofted into the region of cold temperatures where the condensation/growth/decay process cyclically repeats. One signature of this process would be a layer of enhanced H2O lying just below the cloud layer, and AIM measurements of water vapor are being used to investigate the occurrence of such water vapor enhancements (see section 2.6.4). The cycling time for this microphysical process will depend on upwelling rates, which will determine whether particles remain buoyant long enough to grow. Investigation of the PMC “rings” described in Section 2.3.1, which are likely related to mesospheric convection, is providing additional information allowing this fundamental hypothesis of PMC formation, growth, and decay to be tested.

Applied to studies such as these, the AIM measurement complement is poised to address outstanding issues regarding PMC microphysics. Using statistical correlations between PMCs and H2O and temperature, we will isolate which of the two, if either, is the key driver for cloud formation. We will estimate the amount of water taken up in clouds and compare this with the measured cloud densities and particle sizes.

The CIPS UV imagery has provided exceptional data on the structure and variability of PMCs at high, summer-time latitudes. A key result is that the PMCs appear to be much more patchy than previously thought, exhibiting extensive “cloud-like” features. While a number of strong gravity wave events have been observed within the polar region they do not appear to be as numerous as wave events observed at lower latitudes (as determined from prior ground-based NLC and airglow measurements). Instead, the CIPS imagery shows a mixture of wave activity and large convective-like features (which may also be related to the underlying gravity wave activity). This is illustrated in Figure 2.6.2-1 which shows a processed CIPS image containing a well-defined wave event (horizontal wavelength 45 km) with many coherent crests extending over 1000 km in length. This event is evident in the upper left of the image, with a distinct boundary. Elsewhere in the scene the PMC are dominated by large dark circular structures that AIM has shown to be a common feature of PMCs.

Our initial research has focused on characterizing the properties of the most prominent wave events (Chandran et al., 2008), which comprise a variety of medium and small-scale wave events (horizontal scales typically 50-300 km), often extending over exceptionally large areas. Planned research includes:
1. In-depth investigation of occurrence frequency and latitudinal dependence of gravity waves in the northern polar cap region to determine their summer-time climatology for the first time.
2. Measurement of the occurrence and properties of convective-like features and investigation of their potential association with gravity waves and lower atmospheric sources.
3. Novel inter-hemispheric comparison of gravity wave occurrence and characteristics at polar latitudes using new southern-hemisphere PMC data.
4. Exploratory investigation of gravity waves at lower latitudes using their stratospheric ozone signal in the CIPS data.
5. Modeling the effects of the observed gravity wave events on the nucleation and growth and/or dissipation of PMCs

We are beginning to understand the dynamics that controls the occurrence of clouds in the high latitude summer mesopause region. We have identified wave modes in the CIPS data that compare favorably with waves seen in the SABER temperature fields [Merkel et al., 2007]. We have developed a model (NOGAPS-ALPHA) which can validate the temperature and H2O fields observed by SOFIE, simulate the waves observed by CIPS and can hindcast the locations of the bright clouds seen by CIPS. One important next step is to quantify and understand the variation in cloud occurrence, brightness and properties between the northern and southern hemispheres. This is a key part of Objective 3 as defined in the original AIM proposal. Since the proposal was written new results have confirmed that interhemispheric differences exist; the AIM data will allow us to better quantify those differences and to understand them for the first time.

We are just now collecting data for our first southern summer. However, even at this early point, important differences have emerged between the two hemispheres. In general, SH clouds are dimmer and less frequent. This is summarized in the SOFIE data shown in Figure 2.6.3-1. The left hand figure is current data for the NH (top row: cloud altitude, bottom row: occurrence frequency) while the right hand column shows data from the 2007 SH.

Whereas clouds are practically ubiquitous in the NH data, with an occurrence frequency greater than 95% at solstice, in the SH, the occurrence frequency hovers in the 70-90% range. For the brightest clouds, this difference is greater; there is up to a 3X greater chance of seeing a bright cloud in the NH. This is also reflected in the CIPS imagery. We are just now beginning to compare the albedo differences seen by CIPS with the frequency and brightness differences seen by SOFIE.
Furthermore, with the detailed composition measurements of SOFIE, we can put these differences on a physical basis. Preliminary data suggest that SH clouds are 3 km higher than in the NH (this is a greater difference than suggested from ground based measurements) and take up less overall H2O. The region from 80-90 km appears to be several degrees K significantly warmer in the SH, relative to the NH (see Figure 2.6.3-2). This is consistent with HALOE (Hervig and Siskind, 2006), but with the more precise SOFIE data and more consistent polar sampling, we will be able to better study these differences throughout the season.

Further, unlike the HALOE data which had to be averaged over 11 years, with each year in the AIM extended mission, we will be able to build up a true year-by-year climatology of these differences. This is critical because there is currently great debate within the community as to the cause of the temperature difference between the NH and SH. Possibilities such as the orbital eccentricity [Lubken and Berger, 2007], differential filtering of gravity waves [Siskind et al., 2003], and teleconnections to the winter hemisphere [Karlsson et al., 2007] have all been raised. We will address these questions with three dimensional models of the atmosphere. The AIM data sampling pattern is identical in latitude for both hemispheres (though different in local time) and nearly identical form year to year, ideal for evaluating trends and hemispheric differences.

PMC particles are composed of ice that forms through condensation of water vapor under supersaturated conditions. Thus understanding the sources and sinks of water vapor is foundational to understanding the formation and evolution of PMCs. The objective of the Hydrogen Chemistry investigation is to understand the abundance and variability of water vapor along with the traces gases which are related to water vapor through chemical and dynamical processes. Study of HALOE water vapor measurements in the polar summer regions led to discovery of a narrow water vapor layer coincident with the PMC layer (Summers et al., 2001). This layer has important implications for the formation and growth of PMC particles and may be an important diagnostic for water vapor sequestering by ice particles.

Figure 2.6.4-1 shows that initial results for SOFIE water vapor confirm an enhancement of water vapor within and near the bottom of the PMC layer as observed in the HALOE data. This layer develops simultaneously with the PMC layer as shown by Hervig et al. (2003) for HALOE data. The dramatically improved S/N of SOFIE over HALOE, along with SOFIE’s better vertical resolution, provide a much more accurate picture of water vapor and its connection to PMCs than was possible with HALOE data. SOFIE measurements of water vapor reveal a smaller and narrower peak in the water vapor layer near 82 km than that seen by HALOE (a 12 year climatology of HALOE measurements). These differences in the layer’s peak value and thickness persist throughout the PMC season. In monthly averages where the HALOE peak values typically exceed 10 ppmv and occasionally reaches 13 ppmv during July and August, the SOFIE water rarely exceeds 8 ppmv. These differences may be explained by the required averaging of HALOE measurements that resulted in denser thinner clouds that vary with altitude being averaged into thicker less dense clouds. The HALOE 6.6 µm location of the HALOE water channel also endured spectral interference from ice and lower S/N, both factors likely producing positive biases in retrieved water results.

We believe the formation of the enhanced water layer is due to sedimentation of PMC particles which is then followed by sublimation near the base of the PMC layer (see Section 2.6.1). Taken at face value, the lower values of water vapor in the PMC layer as seen by SOFIE data imply significantly less sequestration of water than was implied by HALOE. The vertical width of the water layer in SOFIE data is very narrow, approximately 3-5 km max, whereas the HALOE layer was broader by more than a factor of two, probably due to the averaging and lower vertical resolution. Also, HALOE showed an indication of a second, albeit smaller, layer near 75 km. There is no evidence of this second layer in the SOFIE data. SOIFE is leading to an improved understanding of the HALOE data analysis.

Changes in the amount of water vapor within the PMC layer near 82 km might be driven by changes in residual circulation (Karlsson et al., 2007), solar UV, temperature (DeLand et al., 2003), or possibly even diurnal effects. Any of these could influence the PMC water vapor layer. If the peak water vapor is connected to solar cycle effects, then future years may exhibit higher values of water vapor as seen in some years of the HALOE observations. Also, SOFIE and HALOE sampling is not identical, so diurnal effects may explain some of the differences in water vapor seen by the two instruments. Current validation efforts are underway to understand these differences. Whether the differences are due to better SOFIE S/N, changes in mesospheric circulation, or perhaps diurnal effects, the new SOFIE data will have important implications for understanding the processes which control the water vapor sequestration by PMC particles.
Photolysis of mesospheric water vapor initiates the catalytic odd-hydrogen (HOx) destruction of ozone in the upper mesosphere. Recently published results from HALOE [Siskind et al., 2007] suggest that ozone is enhanced above the PMC cloud layer. This might be due to dehydration and the expected anticorrelation between odd oxygen (Ox) and odd hydrogen chemistry. However, in the Siskind et al. [2007] study the effects of PMCs on temperature were shown to depend upon several poorly understood model assumptions. The use of HALOE temperatures in that study was hindered by the fact that the HALOE temperature retrieval becomes less valid above 80 km where PMC’s form. SOFIE water vapor and ozone data are of vastly better quality than the HALOE data and thus have the potential of significantly improving our understanding of the processes which control mesospheric ozone and its variation.

Most microphysical studies argue that meteoric smoke is the primary PMC nucleation site by default. This is because homogeneous and ion nucleation of PMCs are highly improbable under mesospheric conditions [e.g. Witt, 1969; Gumbel, 2003]. A set of measurements relevant to the topic of cloud nucleation is that from the Cosmic Dust Experiment (CDE) on AIM. CDE determines the magnitude and characterizes the temporal and spatial variability of the cosmic dust influx, allowing for direct correlation studies with PMC frequency and brightness.

As CDE data accumulates we expect to be able to better identify the spatial and temporal variability of the dust influx (Section 2.2.3). A preliminary comparison of a model based on radar observations to the CDE measurements of the latitude dependence of the dust influx is shown in Figure 2.2.3-1. The absolute number of hits per day is qualitatively similar and we are studying the differences as more data becomes available.
The resultant distribution of smoke from ablating meteors is critical in quantitatively understanding PMC microphysics. However, smoke has only been measured below 20 km [Murphy et al., 1998] and its vertical distribution is only modeled above this altitude [e.g. Hunten et al., 1980]. As a result, smoke concentrations are uncertain to several orders of magnitude, which lead to uncertainties of at least a factor of two for modeled cloud brightness and particle radius [Rapp and Thomas, 2006]. Furthermore, recent work has shown that coagulating smoke to sizes large enough for PMC nucleation (~1 nm) is slow compared to meridional transport so that smoke concentrations in the polar summer are extremely low [Megner et al., 2008]. This has only heightened interest in the problem.
We are currently studying the SOFIE data for evidence of smoke particles in the mesosphere. This was never considered possible when casting the original AIM science goals but the unprecedented sensitivity of SOFIE may allow for this detection. The weak smoke signal can be compromised by the presence of PMCs, so we are focusing on the SOFIE data from the winter hemisphere. The challenge in isolating the smoke signal is in distinguishing its weak extinction from the Rayleigh scattered signal in the upper mesosphere. A positive smoke detection would be the first of its kind in the mesosphere and provide new and valuable constraints to models of PMC formation.

Our original proposal asked: What is needed to establish a physical basis for the study of mesospheric climate change and its relationship to global change? This objective is the most far-reaching goal of AIM, and thus has the longest time-table, perhaps requiring data from all four seasons of the nominal AIM mission. The necessary milestones are: (1) validation of the AIM instruments, and (2) establishing causal links between PMC properties and the forcing variables.

Validation: An important milestone in validation of CIPS with the well-calibrated SBUV experiments (three of which are operating contemporaneously with AIM) has already been accomplished, as discussed in Sec. 2.6.1. The continuous time series of data from eight identical SBUV experiments indicate a clear increase of PMC brightness and frequency over the past three decades [DeLand et al., 2007]. In addition, the SOFIE extinction measurements are being compared to HALOE measurements (Section 2.3.3). Validation of CDE is more challenging since this is the first spatially and temporarily resolved measurement of the dust influx. However, the total average influx reported by CDE shows a good agreement with existing models based on a variety of sources, including in situ measurements by the Long Duration Exposure Facility (LDEF), and remote sensing radar observations (see Figure 2.2.3-1).

Establish causal links: This involves establishing statistical relationships between PMC properties and the forcing atmospheric variables, which are believed to be temperature, water vapor and cloud seed nuclei [Thomas, 1991]. SOFIE and CIPS measure both PMC and the relevant atmospheric variables in a common volume. The AIM data set, plus the global TIMED SABER measurements of water vapor and temperature, provide for the first time, the data needed to establish these causal links. Another tool essential to this endeavor is a comprehensive model that includes coupled chemistry, dynamics and ice microphysics. Dynamical transport of ice particles between their genesis in the cold polar regions and their appearance at lower latitudes may be a critical element in this problem [Berger and von Zahn, 2007]. Two versions of the WACCM GCM, combined with detailed ice nucleation, growth, bulk motion, and sublimation are now available for detailed simulations of both the dynamics and microphysical processes [Marsh et al, 2007; Bardeen et al., 2007].