If you look long enough...
 We have investigated the 5 day wave in both temperature and water vapor in the stratosphere and mesosphere as seen in the Navy Operational Global Atmospheric Prediction System–Advanced Level Physics High Altitude (NOGAPS-ALPHA) analysis fields for summer 2007. We have compared these fields and the derived saturation ratios with polar mesospheric cloud (PMC) measurements from the AIM satellite. We find that the 5 day wave is variable in both time and space, with significant amplitudes in the temperature wave in August (up to ∼6 K). By contrast, the 5 day wave–induced water vapor anomalies remain at a near-constant level throughout the season. During August, the 5 day wave in the NOGAPS-ALPHA saturation ratio and in the occurrence of clouds in the AIM data shows a clear anticorrelation with bright PMCs forming in the trough of the temperature wave. The analysis shows that the August enhancement in the 5 day wave amplitude acts to extend the PMC season past the time when zonal mean temperatures are saturated with respect to ice. The increased wave amplitude in August is attributed to in situ wave generation and amplification due to baroclinic instability of mean winds at around 0.1–0.01 hPa. The late-season extension of cloud occurrence due to the 5 day wave may explain previous ground-based reports of bright noctilucent cloud displays in August.
 Planetary Rossby waves with typical periods ranging from ∼2–20 days can be categorized either as forced waves, generated by orographic forcing, strong tropical convection, or longitudinal variations in heating [e.g., Salby, 1984; Holton, 1992], or as free modes, which are resonant responses to atmospheric disturbances. Rossby waves are embedded in Laplace's tidal equation, with the gravest mode being the westward propagating 5 day wave of zonal wave number 1 [Forbes, 1995]. The fundamental theory of planetary waves is well developed [e.g., Ahlquist, 1982; Salby, 1984; Andrews et al., 1987; Volland, 1988; Holton, 1992; Forbes, 1995] and observations have confirmed the basic properties. However, observations also show significant dynamical variability that is only partially understood.
 The 5 day wave is well documented in the lower atmosphere with strong signatures in pressure and geopotential height [e.g., Madden and Julian, 1972; Madden, 1978; Ahlquist, 1982; Speth and Madden, 1983; Lejenäs and Madden, 1992]. In the mesosphere it has traditionally been observed in winds measured by ground-based radar systems [e.g., Lieberman et al., 2003; Riggin et al., 2006] or in temperatures measured from satellites [e.g., Hirooka, 2000; Garcia et al., 2005]. Rosenlof and Thomas  utilized ozone measurements from the Solar Mesosphere Explorer satellite to establish the presence of the 5 day wave in the lower mesosphere, and most recently, Sonnemann et al.  reported a quasi 5 day signal in lower-mesospheric water vapor mixing ratios. Other planetary wave modes with periods of around 2, 10, and 16 days are also often observed in the mesosphere [e.g., Palo and Avery, 1996; Azeem et al., 2001; Lawrence and Jarvis, 2003; Pancheva and Mitchell, 2004; Turnbridge and Mitchell, 2009].
 In the polar summer mesosphere, planetary waves can potentially modulate high-altitude clouds, commonly referred to as noctilucent clouds (NLCs) by ground-based observers, and as polar mesospheric clouds (PMCs) when observed from space. While ground-based observations have revealed planetary wave modulation of PMCs with periods of 5 and 16 days [Kirkwood et al., 2002; Kirkwood and Stebel, 2003], satellite measurements of PMCs have provided us with an opportunity to study the evolution and dynamics of the 5 day wave and its impact on the hemispheric PMC field [Merkel et al., 2003, 2008, 2009; von Savigny et al., 2007]. A westward propagating wave number 2 mode with a period near 2 days has also been documented in the PMC field [Merkel et al., 2008, 2009] which, together with the 5 day wave work of Kirkwood et al.  and von Savigny et al. , show that planetary wave modulations of PMC properties are strongly coupled to planetary wave temperature oscillations.
 The NASA Aeronomy of Ice in the Mesosphere (AIM) satellite is dedicated to the study of PMCs. AIM was launched into a sun synchronous orbit at 600 km altitude on 25 April 2007 with the primary goal of investigating how PMCs form and vary. To attain this goal, AIM carries three instruments: the Solar Occultation For Ice Experiment (SOFIE), which is a solar occultation instrument, the Cloud Imaging and Particle Size (CIPS) experiment, which is a panoramic UV imager, and finally, a dust collector named the Cosmic Dust Experiment (CDE). An overview of the AIM mission and its instruments is provided by Russell et al. . A relatively strong 5 day wave in early August was noted by Benze et al.  using PMC occurrence rates from the second generation Solar Backscatter Ultraviolet Instrument (SBUV/2) and CIPS. A more detailed spectral analysis of the CIPS albedo showing 5 day wave–induced cloud variability has been described by Merkel et al.  utilizing CIPS imagery from 1 June to 15 August 2007. On comparing these data with Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) temperature measurements, they concluded that most clouds formed within the cold phase of the temperature wave. However, the phase difference between 5 day signals in PMCs and temperature varied between 150° and 180°, supporting earlier inferences by Merkel et al. . Merkel et al.  suggested that this varying phase difference could be attributed to the net effects on PMCs of differently phased 5 day wave responses in temperature and water vapor mixing ratios.
 In support of the AIM mission, scientists at the Naval Research Laboratory have run their prototype high-altitude global numerical weather prediction (NWP) system, known as the Navy Operational Global Atmospheric Prediction System with Advanced Level and Physics and High Altitude (NOGAPS-ALPHA). This system was used to assimilate satellite measurements of stratospheric and mesospheric temperature, water vapor, and ozone during the first AIM PMC season covering the period 15 May–31 August 2007. These ground-to-mesosphere NOGAPS-ALPHA analysis fields provide a global synoptic view of planetary-scale dynamics that can affect PMC formation. Hoppel et al.  described an initial NOGAPS-ALPHA experiment that assimilated high-altitude satellite temperature observations in January 2006 up to 0.01 hPa. Eckermann et al.  extended the system to assimilate temperature, water vapor, and ozone observations up to 0.0022 hPa during May–July 2007.
 Eckermann et al.  studied the planetary wave signals in these analysis fields near the polar summer mesopause using two-dimensional space-time Fourier analysis techniques, revealing the presence of several large-scale waves, with the most prominent being the 5 day wave (westward wave number 1), consistent with the observational findings of Merkel et al. . Other significant wave modes were the migrating diurnal and semidiurnal tides, and a westward propagating wave number 2 mode with a period near 2 days. According to Eckermann et al. , the 5 day wave reached its peak amplitude in temperature near the 0.01 hPa level (geometric altitude ∼80 km) at 30°N during the month of June. Furthermore, the peak region was extensive, spanning the 20°N–70°N latitude band, with the peak at 70°N occurring at ∼77 km altitude. Zonal mean 5 day wave amplitudes showed considerable variability, with major peaks occurring 20 days prior to and after solstice. The largest amplitudes appeared in the 50°N–55°N band. However, significant peak activity was also evident at higher latitudes of 70°N–85°N.
 Eckermann et al.  also investigated the 5 day wave signal in water vapor mixing ratios and found a peak near 60°N–75°N with amplitudes ∼0.2–0.3 ppmv, consistent with ground-based measurements at 69°N by Sonnemann et al. . The subsequent spectral cross-coherence analysis between the wave signals in temperature and water vapor, designed to test the hypothesis of Merkel et al.  (that 5 day wave signals in both temperature and humidity control observed 5 day modulation of PMCs), showed no statistically significant correlation at ∼5 days. Eckermann et al.  attributed this to two possible factors. First, large measurement errors in the assimilated satellite water vapor observations at these altitudes might prevent small 5 day wave–induced water vapor anomalies from being resolved with sufficient accuracy. Second, the reported 5 day wave perturbations in temperature exhibited typical amplitudes, and these moderate wave perturbations may not be sufficiently large to generate responses in the analyzed humidity fields.
 Merkel et al.  and Eckermann et al.  showed that the most dominant planetary wave mode during the 2007 northern hemisphere PMC season was the westward propagating 5 day wave number 1 mode. In this study we investigate its variability, origin, and the phase relations among 5 day wave anomalies in temperature, water vapor, and saturation ratio and observed PMC variability. We utilize the NOGAPS-ALPHA gridded analysis fields covering the full first AIM PMC season from 15 May to 31 August 2007, which extends beyond the periods studied by Merkel et al.  and Eckermann et al. . The synoptic fields provide an opportunity to study the 5 day wave in the polar mesosphere in greater temporal and spatial detail utilizing wavelet techniques. As satellite temperature and water vapor observations are assimilated up to 0.0022 hPa, we use the gridded analysis fields and CIPS observations of PMC occurrence frequencies to investigate the phase relations among 5 day wave signals in PMCs, temperature and water vapor at PMC altitudes, to test the hypothesized role of water vapor in affecting PMC brightness modulation by the 5 day wave [Merkel et al., 2009]. Furthermore, we investigate the dynamical impact of the 5 day wave on bright PMCs observed by SOFIE.
 Section 2 describes the NOGAPS-ALPHA assimilation system and the specific wavelet techniques applied in this study. Section 3 describes the 5 day wave signatures in NOGAPS-ALPHA temperature, water vapor, and derived saturation ratio, and how they relate to SOFIE and CIPS cloud observations. Section 4 investigates a significant enhancement in the 5 day wave amplitude during early August and its impact on PMC observations at high latitudes in the late PMC season. Major findings are summarized in section 5.
2.1. NOGAPS-ALPHA Meteorological Analysis
 Here we use 6 hourly global analysis fields at geometric heights z ∼0–90 km generated by an Advanced Level Physics High Altitude (ALPHA) prototype of the Navy Operational Global Atmospheric Prediction System (NOGAPS). Eckermann et al.  describe the configuration of the NOGAPS-ALPHA forecast model and data assimilation system used to perform the specific forecast assimilation runs whose analysis output is analyzed here. Briefly, the spectral forecast model was run at T79L68, yielding a quadratic Gaussian grid resolution of ∼1.5° and vertical resolution in the stratosphere and mesosphere of ∼2 km extending to 0.0005 hPa. Parameterized nonorographic gravity wave drag was tuned to reproduce observed