ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-18-8893-2018Local time dependence of polar mesospheric clouds: a model studyLocal time dependence of polar mesospheric cloudsSchmidtFrancieBaumgartenGerdbaumgarten@iap-kborn.dehttps://orcid.org/0000-0002-6727-284XBergerUweFiedlerJensLübkenFranz-JosefLeibniz-Institute of Atmospheric Physics, Rostock University, Kühlungsborn, GermanyGerd Baumgarten (baumgarten@iap-kborn.de)26June201818128893890818August201711October201730May201814June2018This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/This article is available from https://acp.copernicus.org/articles/18/8893/2018/acp-18-8893-2018.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/18/8893/2018/acp-18-8893-2018.pdf
The Mesospheric Ice Microphysics And tranSport model (MIMAS) is used to study
local time (LT) variations of polar mesospheric clouds (PMCs) in the Northern
Hemisphere during the period from 1979 to 2013. We investigate the tidal
behavior of brightness, altitude, and occurrence frequency and find a good
agreement between model and lidar observations. At the peak of the PMC layer
the mean ice radius varies from 35 to 45 nm and the mean number density
varies from 80 to 150 cm-3 throughout the day. We also analyze PMCs in
terms of ice water content (IWC) and show that only amplitudes of local time
variations in IWC are sensitive to threshold conditions, whereas phases are
conserved. In particular, relative local time variations decrease with larger
thresholds. Local time variations also depend on latitude. In particular,
absolute local time variations increase towards the pole. Furthermore, a
phase shift exists towards the pole which is independent of the threshold
value. In particular, the IWC maximum moves backward in time from 08:00 LT
at midlatitudes to 02:00 LT at high latitudes. The persistent features of
strong local time modulations in ice parameters are caused by local time
structures in background temperature and water vapor. For a single year local
time variations of temperature at 69∘ N are in a range of ±3 K
near 83 km altitude. At sublimation altitudes the water vapor variation is
about ±3.5 ppmv, leading to a change in the saturation ratio by a factor
of about 2 throughout the day.
Introduction
Polar mesospheric clouds (PMCs), also known as noctilucent clouds (NLCs),
consist of water-ice crystals. They occur at midlatitudes to high latitudes, around
83 km altitude e.g.. Such
clouds form in summer in a supersaturated cold atmosphere with temperatures
below 150 K and are sensitive to water vapor and mesospheric temperatures.
Therefore, PMCs are thought to be sensitive indicators of climate change in
the middle atmosphere
e.g.. PMCs often show a
rich variability which provides information about thermal and dynamical
processes on thermal background fields . The clouds have
been shown to be subject to persistent local time variations
e.g.. These
variations were attributed to atmospheric thermal tides. Such tidal
oscillations are globally forced due to absorption of solar irradiance
throughout the day. While semidiurnal tides are dominantly generated through
absorption of solar ultraviolet radiation by stratospheric ozone, water vapor
in the troposphere absorbs solar radiation in the near-infrared bands, mainly forcing
the diurnal tidal component . Generally, these
tidal waves propagate upwards with exponential growth in amplitude and are
therefore also present at PMC altitudes in the summerly mesopause region at
high latitudes.
A variety of spaceborne experiments have observed PMCs since the late 20th
century e.g.. Many of
these experiments are on satellites with sun-synchronous orbits and therefore
only allow observations at fixed local times. The Solar Backscattered
Ultraviolet (SBUV) instruments onboard the National Oceanic and Atmospheric
Administration (NOAA) satellites provide a data set of more than 35 years of
PMC observations e.g.. This data set was recorded by
eight separate instruments with changing viewing conditions and different
local times, which introduces uncertainties in the long-term analysis when
creating a single data set. Also, the Solar Occultation For Ice Experiment
(SOFIE) and the Cloud Imaging and Particle Size (CIPS) instrument onboard
the Aeronomy of Ice in the Mesosphere (AIM) satellite perform observations in
a sun-synchronous orbit. The Ozone Monitoring Instrument (OMI) onboard the
Aura satellite is able to measure PMCs at different local times, but only part
of the diurnal cycle is covered, i.e., the afternoon is missing
. In order to quantify long-term natural or
anthropogenic changes in PMCs, it is therefore essential to understand their
variations over the diurnal cycle .
In contrast to satellite measurements, ground-based measurements are
geographically restricted but have the ability to cover a full local time
cycle. For example, variations of PMC occurrence frequency and brightness as a function
of local time have been observed in detail with lidar instruments
. All these data show evidence of a large PMC brightness
variability with local time.
In this paper we discuss results from a 3-D Lagrangian transport model for
PMCs called MIMAS (Mesospheric Ice Microphysics And tranSport model); see
also the data description in . MIMAS covers the
latitude and altitude range of PMCs and the entire PMC season with a high
temporal resolution. This allows, for example, the calculation of
latitude-dependent local time adjustments to retrieve PMC parameters with the
observational filter of satellite instruments. In the next section we
describe some important aspects of the MIMAS model which are relevant for the
simulation of seasonal and local time variations in PMCs. Here, we also
describe some mean atmospheric background conditions and we characterize
local time variations of background temperature and water vapor as calculated
by the model. Furthermore we give an overview of local time variations in
backscatter (Sect. 3), ice water content (Sect. 4), ice particle radius,
number density, and ice mass density (Sect. 5) seen in MIMAS and compare
these values to lidar and satellite observations. Finally, we discuss the
latitudinal dependencies (Sect. 6) of local time variations in IWC and their
possible implications when analyzing satellite data at fixed local times.
The MIMAS ice modelModel description
The MIMAS model is a 3-D Lagrangian transport model designed
specifically to model ice particles in the mesosphere–lower thermosphere
(MLT) region. MIMAS is limited from midlatitudes to high latitudes
(45–90∘ N) with a horizontal grid of 1∘ in latitude and
3∘ in longitude, and a vertical resolution of 100 m from 77.8 to
94.1 km (163 levels).
Typically, MIMAS calculates a complete PMC season from mid-May to end of
August. Each of the seasonal simulations starts with the same water vapor
distributions on constant pressure levels . Then, the
background water vapor is transported by 3-D winds, mixed by turbulent
diffusion, and reduced by photo-dissociation from solar ultraviolet
radiation. We use Lyman-α as a proxy for solar activity (available at
http://lasp.colorado.edu/lisird/lya/, last access: 2016).
Simultaneously, 40 million condensation nuclei (dust particles) are
transported according to 3-D background winds, particle eddy diffusion, and
sedimentation. The radii of the dust particles in the model vary according to
a Hunten distribution between 1.2 and 3.6 nm . While
each of the 40 million particles is transported on an individual 3-D
trajectory with a time step of 45 s, a single dust particle will nucleate or
an already existing ice particle will further grow, respectively, whenever
the temperature and water vapor concentration of the background atmosphere
provide conditions of supersaturation. In the case of undersaturated
conditions a preexisting ice particle will start to sublimate. The local
formation, growth, and sublimation of all ice particles are interactively
coupled to the local background water vapor concentration which leads to a
redistribution of H2O with local freeze drying and water supply
.
In MIMAS, temperatures, densities, pressure, and wind fields are prescribed
using hourly output data from the Leibniz Institute Middle Atmosphere (LIMA)
model which aims in particular to represent the thermal structure around
mesopause altitudes . LIMA is a fully nonlinear,
global, and 3-D Eulerian grid-point model taking into account
major processes of radiation, chemistry, and transport. LIMA extends from the
ground to the lower thermosphere (0–150 km) and applies a triangular
horizontal grid structure with 41 804 grid points in every horizontal layer
(Δx≈Δy≈ 110 km). This allows the resolution of the
fraction of the large-scale internal gravity waves with horizontal
wavelengths of ≥ 500 km.
Local time variation derived from monthly and zonal means of
temperature (a) and water vapor (b) in the latitude band
67–71∘ N for July 2009; see text for more details.
LIMA is nudged to tropospheric and stratospheric reanalysis data available
from the European Centre for Medium-Range Weather Forecasts (ECMWF), Reading,
UK. LIMA incorporates the 40-year ECMWF reanalysis data set (ERA-40) from
1960 to 2002 and ECMWF operational analysis thereafter. The nudging
coefficient is altitude dependent with a constant value of
1 (3.5 day)-1 from the ground to the middle stratosphere (35 km).
Above 35 km, the coefficient linearly decreases to zero until 45 km. The
nudging of ECMWF data introduces short-term and year-to-year variability.
Above approximately 40 km, carbon dioxide and ozone concentrations as well
as solar activity vary with time. For CO2 we have used a monthly
mean time series for the entire period (1961–2013) as measured at Mauna Loa
(from http://www.esrl.noaa.gov/gmd/ccgg/trends, last access: 2016). For
ozone, we take a temporal variation in the height region of the upper
stratosphere and lower mesosphere (40–65 km) into account. More precisely,
we have used relative anomalies at 0.5 hPa from 1979 to 2013 as measured by
SBUV satellite instruments (from
https://acd-ext.gsfc.nasa.gov/Data_services/merged/index.html, last
access: 2016); for more details see . Before 1979 ozone
data are taken from the World Meteorological Organization (WMO) report
. Finally, daily Lyman-α fluxes from January 1961
until December 2013 are taken as a proxy for solar activity.
Mean state and local time variations of atmospheric background temperature and water vapor
Background conditions of temperature and background water vapor
are certainly of overriding importance, controlling ice formation in the mesopause
region. In the following we will briefly summarize some of the main MIMAS results of
mean state and local time variation of the background. We show in
Fig. examples of monthly and zonally averaged
temperature and water vapor fields as a function of local time in the
Northern Hemisphere for July 2009. We choose the altitude region 81–84 km
in order to resolve typical background conditions of temperature and water
vapor concentrations at PMC heights. We selected a single year, namely 2009,
to be unaffected by possible long-term variations of the local time behavior.
In addition, the year 2009 was analyzed in detail by previous studies
. The monthly average shown in
Fig. has been determined using an hourly output of
temperatures and water vapor from the MIMAS 1∘× 3∘× 100 m latitude–longitude–height grid. For each hourly data set, the actual
longitudinal position on a latitudinal circle is transformed to a uniform
local time. Hence, our local time resolution is defined by the number of 120 longitudinal grid points for a given hourly data set. Finally, we calculate
the monthly July average from 31 (days) times 24 samples per day. We note
that this averaging process resolves the mean sun-synchronous part of
migrating tidal oscillations. In the following we name this procedure
“method 1”, which allows the identification of mean local time variations based on a
monthly zonal average.
Local time variation derived from daily data of temperature (K) for
two heights (km) at different latitudes for July 2009; see text for more
details. Mean: mean temperature over a daily cycle; Max.: maximum temperature
over a daily cycle; Min.: minimum temperature over a daily cycle; LT(Max.):
local time (LT) in hours of Max.; LT(Min.): local time (LT) in hours of Min.; A24:
diurnal amplitude from a harmonic fit including 24 and 12 h components;
A12: same but for the semidiurnal amplitude; P24: diurnal phase of
A24 in LT hours of maximum; P12: same but for semidiurnal phase.
Another possibility to examine local time structures is to analyze
straightforward time series of a single day based on hourly data for
individual latitudinal and longitudinal grid points (“method 2”). We then
estimate from each daily data sample specific parameters of mean and
maximum/minimum values including corresponding times. Additionally,
sinusoidal fits are applied to this daily sample in order to calculate 24,
12, and 8 h tidal amplitudes and phases. This procedure is repeated for
every grid point, taking into account the difference in local time on various
longitudinal positions, and for every day during July. After averaging, we
finally get mean values of parameters that describe monthly local time
variations on the basis of local daily time series. Generally, method 1
generates smaller estimates of mean local time variations than method 2 since
local time parameters are determined from a highly smoothed state in method 1. Conversely, method 2 uses single-day time series and therefor also
records day-to-day variations of daily fluctuations which depend not only on
variable tidal wave activity but also on variable planetary and large-scale
gravity wave activity, e.g., as observed by .
However, our MIMAS simulations are driven by hourly inputs and not by a
monthly zonal mean state. For this reason results from method 2 should better
describe mean local time fluctuations of background conditions that effect
ice formation. Table summarizes some relevant numbers
that describe the mean state and local time fluctuations of temperature
resulting from method 2.
We begin with a short discussion of the general mean background state of
temperatures. Both averaging procedures from method 1
(Fig. ) and 2 (Table , third
column) result in identical monthly mean values of temperatures. The modeled
temperatures closely match observed mesopause temperatures and altitudes.
Monthly mean MIMAS temperatures at 69∘ N are very similar to the
observed temperature climatology derived from rocket (falling spheres)
measurements at ALOMAR (69∘ N) during summer
. For example, the minimum temperature is
∼ 130 K in MIMAS compared to ∼ 130 K in the climatological
observations for July, and the summer mesopause altitude is also basically
identical (∼ 88 km). At typical NLC heights at 83 km, mean MIMAS
temperatures are slightly higher, with ∼ 147 K compared to observed
∼ 145 K. The MIMAS summer mesopause at 78∘ N (89 km, 124 K)
is colder and higher compared to lower latitudes. Lidar measurements of
temperatures were performed in the upper MLT at Spitsbergen (78∘ N)
in the years 2001–2003. The July observations show that the summer mesopause
is located at 90 km and is as cold as 122 K . At lower
latitudes at 54∘ N the MIMAS mesopause is significantly lower
(86 km), warmer (144 K), and less pronounced. Again, lidar observations of
temperatures confirm these model results with, for example, a mean July
mesopause (86 km, 147 K) at Kühlungsborn (54∘ N)
. So far, we validated model temperatures of the summer
mesopause region only with observational climatologies obtained from
ground-based lidar facilities and rocket measurements that we think provide
reliable data sets for the high-latitude MLT region. Furthermore,
ground-based measurements with meteor radars indicate low temperatures around
90 km in summer, typically in a range of 150–170 K at 54∘ N and
120–140 K at 69∘ N . Calculated
temperatures from MIMAS fit to these observations; see
Table . (see their Fig. 1) also
published temperatures at 68∘ N (23:06–23:18 LT) for July
2009 observed by the SOFIE satellite instrument which show similar
temperatures compared to rocket measurements. For example, SOFIE temperatures
indicate a mesopause at 88 km with a mesopause temperature of
∼ 135 K.
Figure and Table also show mean
daily temperature fluctuations. Looking at Fig. ,
local time variations calculated with method 1 have a value about
± 1–1.5 K near 83 km at 69∘ N. Applying our preferred
averaging procedure from method 2 yields systematically larger local time
variations; see Table . The analysis shows that in the
height region 83–90 km, local time variations of temperature decrease
towards the pole, i.e., ±5–15 K at 54∘ N,
±3–11 K at 69∘ N, and ±2–6 K at 78∘ N.
Generally, the tidal analysis of temperatures indicates that diurnal and
semidiurnal tides are mainly present, whereas the terdiurnal component can be
neglected. Thermal amplitudes increase with altitude and decrease with
poleward direction, as has been discussed in . Absolute
values of diurnal and semidiurnal amplitudes from MIMAS are in the same order
as has been calculated in the model study by and
. Also, tidal temperature variations derived from meteor
radar observations around 90 km in summer show diurnal (semidiurnal)
amplitudes of about 7 K (5 K) at 54∘ N, and amplitudes of about
4–8 K (2–4 K) at higher latitudes, such as 69∘ N . These
observations match the size of amplitudes estimated by MIMAS; see Table .
At PMC altitudes near 83 km diurnal tidal amplitudes are up to a factor of 2
stronger than semidiurnal amplitudes. This means that local variations of
temperatures are mainly affected by diurnal tidal modes. At mesopause
altitudes diurnal and semidiurnal amplitudes get larger and are of similar
size.
Local time variation derived from daily data of H2O (ppmv)
for two heights (km) at different latitudes for July 2009; see text for more
details. Mean: mean H2O over a daily cycle; Max.: maximum
H2O over a daily cycle; Min.: minimum H2O over a daily
cycle; LT(Max.): local time (LT) in hours of
Max.; LT(Min.): local time (LT) in hours of Min.; A24: diurnal amplitude
from a harmonic fit including 24 and 12 h components; A12: same but for
the semidiurnal amplitude; P24: diurnal phase of A24 in LT hours of
maximum; P12: same but for semidiurnal phase.
We also compared the phase structures as calculated by the two averaging
procedures from method 1 and 2, and find that phases of maximum and minimum
values as well as tidal phases remain almost unchanged. Interestingly,
temperature phases change with latitude at PMC altitudes. Particularly, the
local time of the daily minimum (Table , seventh column)
is shifted backwards in time towards higher latitudes from 06:36 LT
(54∘ N) to 04:24 LT (69∘ N) and 02:00 LT
(78∘ N). Contrary to the shift of the minimum, the time of
temperature maximum seems to occur steadily always between 15:00 and
17:00 LT. The superposition of diurnal and semidiurnal thermal tides causes
predominantly lower temperatures during early morning hours and higher
temperatures during afternoon hours, respectively.
Besides temperatures, water vapor plays an essential role for PMC formation.
Figure shows water vapor mixing ratios from MIMAS ice
simulations at latitudes 67–71∘ N for July 2009. In
addition, Table describes numbers, using method 2, of
latitudinal dependencies for daily variations of water vapor.
At 69∘ N the mean vertical water vapor profile maximizes at 81.5 km with 8 ppmv where ice particles
sublimate and create a zone of enhanced hydration. SOFIE observations of
water vapor at 73∘ N show a similar vertical structure with a water
vapor peak of 8 ppmv near ∼ 83 km . From
Table we find that effects of hydration (sublimation of
ice) at 81.5 km and dehydration (freeze drying) near 84 km are intensified
towards higher latitudes since colder mesopause temperatures permit larger
nucleation rates of ice particles, and larger sedimentation paths lead to
enhanced growth of ice particles that causes enhanced sublimation.
MIMAS results indicate that local time variations of water vapor in terms of
absolute values are much stronger than thermal local time variations. At
69∘ N local time variability of background water vapor can reach
values up to 7 ppmv at 81.5 km, which is in the order of a 100 %
variation. Consequently, tidal amplitudes of water vapor from harmonic fits
show large tidal components with an increase towards higher latitudes
contrary to temperature amplitudes. The local time behavior of water vapor
shows a pronounced maximum below PMC altitudes at 81.5 km during the morning
between 05:00 and 07:00 LT. The phase position of maximum water vapor moves to some
extent backwards in time in the poleward direction, however, with a delay of
approximately 3 h when compared with temperature phases. Hence, both
phase positions of low temperatures and large water vapor mixing ratios
approximately coincide. For this reason we expect that the maximum strength
of PMC formation should occur during morning hours, as we will discuss in the
next sections.
Generally, modeled PMCs in MIMAS exist approximately poleward of
54∘ N, where the degree of mean saturation S is larger than unity. Saturation
conditions are a combined effect of temperature, water vapor, ambient
pressure, particle size, and particle temperature. Figure shows
the saturation ratio S at a fixed altitude of 82.7 km, which is the mean
PMC altitude in the MIMAS simulation for the year 2009. The saturation ratio
S is approximated by S=pH2O/p∞ with equilibrium pressure
p∞ and ambient partial pressure pH2O=c(H2O)⋅p, where
c(H2O) is the volume mixing ratio of water vapor and p is pressure of
air;
for details see Eqs. (1)–(3) in .
It was found that for most of the time that supersaturation exists, the
saturation ratio only falls below S=1 in the afternoon hours. The July average shows nearly permanently supersaturated
conditions throughout the day. Note that the vertical extent of
supersaturation areas increases polewards because of colder and higher
mesopause conditions. In the following sections we will present model results
of different PMC parameters and compare these with observational data.
Hourly mean values of the saturation ratio (S≊pH2O/p∞) in the latitude band 67–71∘ N for July
2009 at a fixed altitude of 82.7 km (mean PMC height) as a function of local
time. Grey lines show individual days and the blue line their mean.
Mean seasonal variations of PMC occurrence
frequency (a, b), altitude (c, d), and brightness
βmax(e, f) between 2003 and 2013 at ALOMAR for
faint (red), long-term (blue), and strong (green) clouds (for details see
text). Panels (a, c, e) show model results for 67–71∘ N,
10–20∘ E; panels (b, d, f) show lidar observations from
ALOMAR. The solid lines represent third-order polynomial fits based on daily
means. Numbers in the figure legends are seasonal mean values. Brightness
ranges for cloud classes are scaled down by a factor of 4 for MIMAS results.
Note the different scaling of the brightness axis for model and lidar data.
Comparison of MIMAS backscatter model results with ALOMAR lidar observationsSeasonal variation of backscatter
During the northern hemispheric summer PMCs typically occur from end of May
until mid-August e.g.. At the core of the ice season in July, lowest temperatures near
130 K have been observed at mesopause altitudes near 88 km at 69∘ N
. Hence, we expect PMCs to be most frequent and bright during
July.
Figure shows the mean seasonal variations of basic PMC
parameters as calculated by MIMAS and observed by the
Rayleigh–Mie–Raman (RMR) lidar at the Arctic Lidar Observatory for Middle
Atmosphere Research (ALOMAR), located at 69∘ N, 16∘ E
. MIMAS results are limited to a latitudinal and
longitudinal area of 67–71∘ N and 10–20∘ E to be
close to the lidar position. We will use the volume backscatter coefficient
of ice particles βmax, in units of 10-10 m-1 sr-1, as a measure for the cloud brightness. Both model and
observations cover the same time period of 11 years from 2003 to 2013. In
order to take different cloud classes and the detection sensitivity of the
lidar into account, we sort measurements and model results into different
brightness ranges: 1 < βmax < 4 (faint
clouds), βmax > 4 (long-term detection limit
of the lidar), and βmax > 13 (strong clouds)
e.g..
In order to convert the model output from MIMAS to specific lidar
measurements, we apply spherical Mie-theory calculations to modeled ice
particle distributions while taking into account the laser wavelength
(532 nm) and scatter geometry (180∘). Finally, the transformed model
results are sorted into brightness ranges. PMC brightness is proportional to
the number of ice particles and depends approximately on the power of 6 on
ice particle radius. For example, increasing the mean radius by only 25 %
from 32 to 40 nm would result into a brightness change by a factor of 4.
It is this high sensitivity of cloud brightness to particle size that forms a
hard benchmark for our complex ice model simulations. On the one hand, small underestimation
of the mean ice radius will dramatically decrease the brightness, whereas on the
other hand, a small overestimation will enhance the resulting backscatter
signal by orders of magnitude. In order to match the mean occurrence
frequencies of the lidar measurements we decreased the brightness ranges,
defining the cloud classes, for the model results by a scaling factor of 4.
Hence, the modeled occurrence frequencies contain a systematic bias. We think
this deficiency is tolerable since our local time analysis relates to
relative deviations from a mean. The scaling factor will only be used for the
comparisons with lidar data in this and the following section.
Figure a and b show a general good
agreement of modeled and observed PMC occurrence frequencies. We find maximum
values in the long-term and strong cloud classes in mid-July around days
relative to solstice (DRSs) 20–30. Faint clouds observed by lidar occur
earlier in the season than modeled faint clouds. This gives a hint that the
model perhaps underestimates the microphysical process of nucleation in ice
formation which essentially determines the frequency of weak PMCs consisting
of small ice particles. We note that ice nucleation in MIMAS is described by
the concept of critical radius
.
Figure c and d show modeled and observed
PMC altitudes which coincide quite well. Generally, weak PMCs are at higher
altitudes compared to strong PMCs. This altitude separation is caused by two
factors. First, the sedimentation velocities of ice particles depend on their
sizes. Weak PMCs consist of ice particle distributions with smaller mean
radii, typically in a range of 20 nm, whereas strong PMCs consist of larger
mean radii, e.g., 40 nm. As the sedimentation velocity increases with
particle size (mass), larger particles can reach lower altitudes along their
sedimentation path. Secondly, smaller ice particles start to sublimate at
lower temperatures than larger ones due to the Kelvin effect. Thus, the
negative vertical temperature gradient of the atmosphere causes smaller
particles to sublimate at higher altitudes than larger particles. As a result
larger ice particles, causing a higher brightness, are found at lower
altitudes.
Figure e and f show modeled and observed
PMC brightness. Here, the model results are calculated according to a given
brightness range as an arithmetic mean of all brightness values matching the
limits. Again, the model seems to underestimate the beginning and end of the season.
The scaling factor for the brightness ranges leads to lower modeled
brightness values in the different cloud classes. Hence, multiplying the
modeled values with the scaling factor of 4 approximately reproduces the
brightness values observed by lidar.
Mean local time variations of PMC occurrence
frequency (a, b), altitude (c, d), and brightness
βmax(e, f) for July in the period from 2003 to
2013 at ALOMAR for faint (red), long-term (blue), and strong (green) clouds
(for details see text). Panels (a, c, e) show model results for
67–71∘ N, 10–20∘ E, panels (b, d, f) show lidar
observations from ALOMAR. The lines represent the sum of four harmonic fits
using periods of 24, 12, 8, and 6 h to hourly mean values. Numbers in the
figure legends are daily mean values. Brightness ranges for cloud classes are
scaled down by a factor of 4 for MIMAS results. Note the different scaling of
the brightness axis for model and lidar data.
We summarize that the modeled seasonal distributions of occurrence, altitude,
and brightness are fairly consistent with the ALOMAR RMR lidar observations,
especially for July conditions. Therefore we will concentrate our discussion
of model results in the following sections on this core period of the
northern PMC season.
Local time variation of backscatter
PMCs preferentially occur during morning hours which is attributed to thermal
tides of background temperatures in the mesopause region
. In order to validate the structure of local time
variations in MIMAS we compare our model results to observations by the
RMR lidar at ALOMAR and to instruments onboard the AIM satellite. For
comparison to lidar data we will apply a scaling factor of 4 regarding the
brightness ranges, defining the cloud classes as described in the previous
section. As discussed above we will concentrate on the core period of the
northern PMC season and will use only July data (31 days × 24 h) from
MIMAS simulations for the PMC seasons 2003–2013. Tidal structures in the
LIMA model have been discussed earlier by and
.
Ratio of diurnal to semidiurnal amplitudes (A24/A12) of
harmonic fits to the modeled and observed occurrence frequency (OF),
altitude, and brightness. The values are calculated for different cloud
classes (for details see text) for July months in the period from 2003 to
2013 at ALOMAR according to Fig. . Bold numbers mark
values that agree within the relative uncertainty of about 15 %
(confidence level of 95 %).
Figure shows the variation of PMC occurrence frequency,
altitude, and brightness throughout the day for the integrated data set of
July 2003–2013 and brightness classes as defined above. The curves are
superpositions of four harmonic functions with periods of 24, 12, 8,
and 6 h, which are fitted to hourly mean values as described in
. The geographic range is again restricted to the area
around ALOMAR. We find pronounced and persistent features which indicate a
strong influence of tides on PMC parameters. The occurrence frequency
variation over a day is largest for strong clouds both in MIMAS and
observations. Like in the observations, the model results show the highest cloud
occurrence during the morning hours. The local time dependencies of altitude
and brightness are anticorrelated, i.e., on average ice clouds of higher
brightness are found at lower altitudes. In general, a predominant diurnal
oscillation exists in agreement with the lidar observations. The lidar
observations show additionally semidiurnal variations in all three PMC
parameters, which seems to some extent underestimated by the model. On the contrary,
the modeled brightness shows a clear peak in the morning hours around 04:00 LT
that is absent in the observations.
In order to investigate these different structures we calculated the ratios
of diurnal to semidiurnal tidal amplitudes (A24/A12). The values in
Table show that both model and lidar fits have
nearly the same amplitude ratios for a number of cloud parameter and class
combinations. For example, for the long-term brightness the ratios are 1.82
(model) and 1.88 (lidar), meaning that tidal modes are very similar in both
data sets. Thus the phase differences of modeled and observed data,
especially for the semidiurnal modes, (not shown here) are mostly responsible
for the differences visible in Fig. . The superposition of
diurnal and semidiurnal tidal modes yields a stronger morning peak in the
modeled compared to the observed brightness.
In summary, observed local time variations of PMC occurrence and brightness
at ALOMAR are fairly well reproduced by MIMAS.
Comparison of MIMAS ice water content model results with AIM satellite observations
Comparison of PMC brightness values between different instruments is affected
by observational constraints, e.g., viewing geometry, lighting conditions,
temporal overlap, and wavelength. suggested that
integrated ice mass has the advantage of being less dependent on instrumental
setups and thus should be more robust to be used for PMC comparisons.
Therefore we present in this section model results of ice water content (IWC)
that are calculated from the integrated ice mass density over the total
vertical ice column. We analyze the time period 2007–2013 to cover the time
range of the SOFIE instrument onboard the AIM satellite. The IWC is
calculated from all longitudes in the latitude band 67–71∘ N. In
order to resolve tidal structures we subdivide each latitudinal circle into
120 longitudinal segments and sort the model data according to actual local
times at all segments. This method yields a total of 4 latitudes times
120 longitudes times 31 days times 24 h of values for July conditions.
Finally, we average all IWC values corresponding to a certain local time with
a local time resolution of 1 h day-1. The probability density
distributions of all these IWC values show to a high degree an exponential
behavior. Therefore we calculate two different averages (median and
arithmetic mean), in order to characterize a mean ice water content as a
function of local time during July.
In Fig. we compare our IWC model results in terms of median
values with measurements from the CIPS and SOFIE instruments onboard the AIM
satellite for the latitude band 67–71∘ N. The AIM satellite
operates in a sun-synchronous orbit; hence only limited local times are
available . For comparison with model results we take
the different sensitivities of the two AIM instruments (SOFIE, CIPS) into
account. The detection threshold for SOFIE is given as 0.5 g km-2. In contrast to SOFIE, the CIPS instrument is less
sensitive, allowing only IWC events larger than 10 g km-2 to be
detectable . Hence all IWC data sets (MIMAS, SOFIE, CIPS)
are limited to this threshold. We find a good agreement between model results
and the data points from SOFIE and CIPS inside the error bars. Generally, the
modeled IWC has maximum values in the early morning hours between 01:00 and
04:00 LT and lowest values between 16:00 and 20:00 LT. On average the IWC
varies by a factor of about 2 during a day. Interestingly, comparing SOFIE
with CIPS data, the CIPS observation at 23:00 LT does not match the SOFIE
point for midnight conditions. There is a substantial deviation between these
values (SOFIE: 60 g km-2, CIPS: 30 g km-2) that might be due to
some uncertainties in the CIPS threshold. The MIMAS value of 40 g km-2
is right in between the two different satellite observations. Nevertheless,
all three data points coincide within their error bars.
We summarize that the MIMAS model results of PMC ice water content are
compatible to a high degree with the satellite observations.
Hourly median values of IWC from 2007–2013 (July) for
67–71∘ N and IWC threshold of 10 g km-2 as a function of
local time. The vertical bars represent the lower and upper quartile of the
data. The black curve is a harmonic fit to the data with periods of 24, 12,
and 8 h. Data from AIM satellite instruments including uncertainties for the
same time range: SOFIE V1.3 (red – from
http://sofie.gats-inc.com/sofie/index.php, last access: 25 November
2016) and CIPS Level 3c-v4.20 (green – from
http://lasp.colorado.edu/aim/download-data-pmc.php, last access:
5 December 2016) for ascending and descending nodes.
Figure shows again the IWC local time variation for the
latitude band 67–71∘ N, but now without any threshold,
which means that IWC has been frequency weighted and IWC values of zero (no PMCs) are included. This yields an IWC variation over day by a factor of 10
compared to the factor of two when considering the threshold used in
Fig. . The factor of 10 derived from frequency-weighted IWC
is consistent with model results reported by (see
their Fig. 7). Hence, the strength of local time variations is sensitive to
the IWC threshold, meaning that larger thresholds induce smaller local time
variations; see discussion in Sect. 6. The times of IWC maxima and minima
are close to those of occurrence frequency and the brightness as shown in
Fig. . We find the harmonic fit to be highly correlated
to the median values (correlation coefficient of 0.99), meaning that the
local time behavior of IWC medians is almost perfectly represented by the
three harmonics of 24, 12, and 8 h. The fit is dominated by the diurnal
and semidiurnal mode, the terdiurnal mode is of minor importance. The
amplitude ratios are A24/A12=2.66 and A24/A8=5.84.
Hourly median values of IWC from 2007–2013 (July) for
67–71∘ N as a function of solar local time. No threshold has been
applied, IWC values of zero (no PMCs) are included. The vertical bars
represent the lower and upper quartile of the data. The black curve is a
harmonic fit to the data with periods of 24, 12, and 8 h.
Local time behavior of ice particle radius, number, and ice mass density
In the previous sections we compared MIMAS simulations of backscatter and ice
water content with observations in order to show that MIMAS provides
realistic model results. Now we investigate in more detail the local time
variations in different ice parameters as ice particle number density, ice
particle radius, and ice mass density in comparison to backscatter.
Ice parameters at 67–71∘ N calculated from MIMAS
simulations of all July months 2003–2013 for the altitude range near 83 km
where βmax > 0.4. (a, b) Brightness
and ice particle radius. (c, d) Ice mass density and particle number
density. The boxes represent lower and the upper quartiles, median (red
line), and arithmetic mean (green line). The dashed vertical bars indicate
the minimum and maximum values.
Our model simulations of PMCs show that the number of ice particles is largest
at mesopause altitudes between 86 and 89 km, where the highest chance of
nucleation is found. This altitude region serves as a reservoir of small ice
particles. Then, slightly below mesopause altitudes, the MIMAS model predicts
the largest number density of ice particles to fall in the range 500 to 1500 cm-3 (67–71∘ N). The mean radius of ice particles
generally stays below 15 nm, which is usually too small to produce
significant lidar backscatter signals. Due to random diffusive transport
processes a fraction of these small ice particles experiences enhanced
growing. The increase in particle mass enhances downward sedimentation.
Towards lower altitudes the amount of free background water molecules
increases exponentially, since air density increases exponentially. During
their downward sedimentation path the growth of ice is stimulated until an
ice particle reaches an altitude where supersaturated background conditions
change into undersaturation. This is the height where the radius of ice
particles maximizes and thus the highest ice mass densities and largest
backscatter signals occur.
In Fig. we present backscatter, mean ice radius, number
density, and ice mass density at the altitude of maximum backscatter signal,
assuming a threshold of βmax > 0.4, for the
latitude band 67–71∘ N during July. The plots show both
median and arithmetic mean values. Median and arithmetic mean are generally
different, which indicates that the underlying distributions are not
symmetric.
Mean ice radii vary between 35 and 45 nm. These numbers are in good
agreement with AIM–SOFIE observations, which also indicate ice radii of
35–40 nm . Mean ice particle densities fall in the
range 80 to 150 cm-3, which agrees with results from lidar observations
and satellite measurements .
Similar to ice radii, the mean ice mass density increases from the heights
below the mesopause downward, with mean values about 30 g km-3 at PMC
heights. It is interesting to note that the low-altitude boundary of the
backscatter at 69∘ N as simulated by MIMAS indicates a temperature of
150 K which agrees well with the observed temperature of 150 ± 2 K
for the low-altitude boundary of NLCs .
Investigating the local time dependence of ice parameters we find that the
ice number density maximizes in the morning hours between 03:00 and 05:00 LT, which
corresponds with the maxima of ice mass density and βmax.
The mean radius shows a smaller variation with local time and no pronounced
maximum in the morning. This indicates that the local time behavior of
ice mass density is mainly determined by the number of ice particles and less
by the ice particle radius. Our model results are confirmed by AIM
observations which show that an increase in ice mass is significantly
correlated with increasing number densities and less correlated with the size
of ice particles . We mention that model calculations
performed with the 1-D ice model CARMA show some controversial results,
meaning that particle number density has no effect on ice mass and brightness
.
In MIMAS local time dependencies in ice parameters are mainly forced by tidal
variations in background temperature and water vapor, as has been discussed
in Sect. 2.2. Local time dependence of brightness in terms of
βmax with a diurnal maximum near 04:00 LT follow nicely
the temperature structure with a diurnal minimum at 04:00–05:00 LT; see
Table 1. In addition, we find the maximum water vapor to occur between 06:00
and 07:00 LT and hence about 2–3 h after the brightness maximum; cf. Fig. and Table 2.
We conclude that the local time phases in temperature and water vapor are the
main drivers to determine the phase structure in ice parameters.
Latitudinal variations of local time dependence for ice water content
Our numerical simulations indicate that the local time variations of PMCs are
subject to significant latitudinal dependencies. Figure shows
modeled IWC values over latitude for selected local times in July
2007–2013. No threshold was applied and IWC values had been frequency
weighted so that median values include “zero” PMC events. While at 06:00 LT IWC
increases nearly linearly from 60 to 84∘ N, the slopes are
quite different throughout the rest of the day. This indicates that the phase
of the local time behavior changes with latitude. As an example, the time of
IWC maximum changes from the morning hours at midlatitudes to midnight hours
at high latitudes. Figure shows this phase variation in
more detail for different latitude bands. It turns out that (1) the amplitude
of the local time dependence increases in absolute IWC values towards the
pole, (2) the ratio of maximum to minimum IWC decreases towards the pole (see
Table ), and (3) a slight phase shift can be seen
with decreasing latitude: the IWC maximum around midnight near 81∘ N
moves forward in time to 04:00 LT near 63∘ N.
Median IWC values for July 2007–2013 as a function of latitude for
different local times. No threshold has been applied, IWC values of zero (no PMCs) are included. The vertical bars represent the lower and upper quartile
of the data.
Ratios of IWC tidal amplitudes for July 2009 and different latitude
bands. No threshold has been applied, IWC values of zero (no PMCs) are
included. The ratios of maximum to minimum IWC indicate the variability
throughout the day. For details see text.
Latitude bandA24A12A24A8Max. / min.61–65∘ N7.66.012.667–71∘ N2.24.118.373–77∘ N2.14.810.479–83∘ N2.77.86.9
IWC median values at midlatitudes are much smaller (about 100 times) than
those at high latitudes. Therefore we also use the ratio of daily maximum to
minimum IWC values as an additional indicator for local time variations; see
Table . Please note that the ratios are calculated
from median IWC values without any lower threshold; hence, the occurrence
frequency has a large influence on the median value. This is in particular
important at the lowest latitude band (61–65∘ N) where
rather small PMC occurrence frequencies are modeled. For example, assuming an IWC
threshold of 5 g km-2, the PMC occurrence frequency at this latitude band
is only in the order of 5–10 % during July, whereas moving poleward it
increases to about 50 % at 67–71∘ N and 100 % at
79–83∘ N. For this reason results for the lowest
latitude band (61–65∘ N) in Table include enhanced uncertainties.
Diurnal variation of hourly median IWC values for July 2007–2013
for different latitude bands. No threshold has been applied, IWC values of
zero (no PMCs) are included. Dots indicate the data and solid lines are
harmonic fits using periods of 24, 12, and 8 h.
Climatology of local time variations of IWC in units of grams per
square kilometer (g km-2) for three thresholds (IWC > 0,
IWC > 10, and IWC > 40) at different latitude bands for the period
July 1979–2013. Mean: mean daily IWC over a daily cycle; Max.: maximum IWC
over a daily cycle; Min.: minimum IWC over a daily cycle; Max. - Min.:
difference between maximum and minimum IWC; Max. / Min.: ratio between
maximum and minimum IWC; LT(Max.): local time (LT) in
hours of Max.; LT(Min.): local time (LT) in hours of Min.; A24: diurnal
amplitude from a harmonic fit including 24 and 12 h components; A12:
same but for the semidiurnal amplitude; P24: diurnal phase of A24
in LT hours of maximum; P12: same but for semidiurnal phase.
Table also includes tidal amplitude ratios
obtained from the fitting of 24, 12, and 8 h harmonic components. We find
that for the three highest latitude bands the diurnal component is generally
about 2 times larger than the semidiurnal component. This ratio
A24/A12 seems to be fairly independent of latitude. There exists a
terdiurnal component with a strength of about 20 % which decreases in the poleward direction. On average the ratio of daily maximum to minimum IWC
values is about 10 and decreases towards the pole.
Now we investigate the local time structure of IWC and its latitudinal
dependence in terms of different IWC thresholds. In the following, IWC data
are not frequency weighted. Additionally we extend the time period to range
from 1979 to 2013, thereby presenting a 35-year climatology of daily
fluctuations which aims to describe mean local time variations. Such
specifications might be useful for satellite data analysis in order to
perform local time corrections. The results are shown in
Table . The modeled IWC data have been calculated over
three latitude bands used in SBUV trend analysis and for three thresholds
with IWC > 0 g km-2, IWC > 10 g km-2, and
IWC > 40 g km-2. The latter threshold was used in SBUV trend analyses
by and . Both absolute means and
absolute local time variations, expressed here as difference between maximum
and minimum value, increase towards the pole. We find the ratio of maximum to
minimum values, a measure for the relative IWC local time variation, to
increase poleward too. Additionally, IWC ratios decrease with higher
thresholds, e.g., at latitudes 64–74∘ N from 6.6
(IWC > 0) to 2.4 (IWC > 10) and 1.7 (IWC > 40); see Table (seventh column).
Maximum values of IWC occur in general during the early morning hours, whereas
minimum values are present in the afternoon hours. Local times of IWC maximum
and minimum are independent of the selected threshold. There exists a time
shift in latitudinal direction, e.g., at polar latitudes
74–82∘ N the maximum occurs at 02:00 LT for IWC > 40 g km-2, whereas at midlatitudes 50–64∘ N it is shifted
forward in time to 08:00 LT. Recently, reported about model
results of PMC IWC calculations with the NOGAPS-ALPHA model using a 1-D bulk
ice model . The authors show that the IWC is largest at
highest latitudes and yields a morning peak between 05:00 and 07:00 LT and a late
afternoon minimum equatorward of 80∘ N regardless of threshold.
Diurnally averaged IWC values (threshold of 40 g km-2) are near
100 g km-2 and consistent with those calculated by MIMAS. NOGAPS-ALPHA
results of IWC over a diurnal cycle show at 68∘ N a ratio between IWC
maximum and minimum of about 1.5 for a threshold of 40 (see Fig. 6a, b in
), similar to a ratio of 1.7 from MIMAS calculations.
Concurrently, absolute IWC local time variations in NOGAPS-ALPHA increase
towards higher latitudes and are threshold dependent. Again, these features
are confirmed by MIMAS.
Lidar observations of daily variations of mid-latitude NLCs (54∘ N,
Kühlungsborn, Germany) show the highest rates at 05:00–06:00 LT, which is
similar to our model result . In contrast,
published local time observations by the Aura OMI
satellite instrument, which indicates maximum frequency and albedo values at
approximately 09:00–10:00 LT at 70∘ N for the NH 2007 season, with
a smaller amplitude and a slight phase shift to ∼ 08:00 LT at higher
latitudes. Hence, model results from MIMAS deviate to some extent from these
satellite measurements for 2007. Here we refer to some year-to-year
variations of phases in MIMAS (not shown here) which might explain to some
extent these differences.
As shown in Sect. 2.2, phase positions of minimum temperature at PMC
altitudes move to some extent during early morning hours backwards in time in the poleward direction. Also, the phase of the daily water vapor maximum tends to
follow this time shift. We conclude that both temperature and water vapor
phases cause the general early morning hour structure in IWC and its shift
towards higher latitudes.
Generally, the time difference between IWC maximum and minimum is
approximately constant, with 12 h at all latitudes and for all three
thresholds. This indicates that a tidal decomposition of daily data reveals
the significant role of the diurnal tidal oscillation. Indeed, all daily time
series of IWC are approximated to a high degree by harmonic fits of a
dominant 24 h and a minor 12 h component, the ratio A24/A12
varies between 4.5 and 8.8. Hence, semidiurnal fluctuations in IWC are of
minor importance, which again is explained by small semidiurnal tidal
amplitudes in temperature and water vapor. We note that terdiurnal tidal
components are also present. But on average 8 h amplitudes are in the order
of 20 % of 12 h amplitudes and therefore have a negligible impact.
We summarize that these results highlight the importance of taking tidal PMC
variations into account when compiling data sets which are distributed over
latitude and local time. It turns out that for IWC (1) local time variations
depend on threshold conditions, e.g., relative local time variations decrease
with larger thresholds; (2) local time variations depend on latitude, e.g.,
absolute local time variations increase towards the pole; and (3) a phase shift
exists towards the pole which is independent of the threshold value, e.g., the
IWC maximum moves backward in time from 08:00 LT at midlatitudes to 02:00 LT at
high latitudes. The IWC local time behavior presumably exhibits year-to-year
as well as long-term variability which may effect the 35-year mean state given
in Table . However, this needs more detailed
investigations and will be subject of future work.
Conclusions
In this paper we presented a detailed investigation of tidal effects on PMC
occurrence, altitude, brightness, and microphysical properties of ice
particles as calculated by the MIMAS model. As already discussed in several
publications, the interpretation of PMC observations requires a careful
treatment of the local time of the observations even for the investigation of
long-term records . We
have compared our results to observations by ground-based lidar as well as
satellite instruments and find a good agreement when taking into account
instrumental sensitivity and local time of observations. MIMAS reproduces the
local time variations seen by lidar especially well in the core of the PMC
season. PMC simulations for ALOMAR show in the latitude range
67–71∘ N brightness variations throughout the day up to
a factor of 7, while the occurrence frequency varies by a factor of 2 to 16
for faint and strong clouds, respectively. At the peak of the PMC layer the
mean ice particle radius varies from 35 to 45 nm and the mean number density
from 80 to 150 cm-3 throughout the day. All quantities show the maximum
around a local time of 3 ± 2 h. At the same latitude band the time of
maximum IWC is about 03:00 LT and the minimum is found around 18:00 LT. Without
thresholding the data, hourly IWC median values vary by a factor of 10
throughout the diurnal cycle. In general diurnal and semidiurnal tides in
temperature and water vapor contribute to the tidal behavior of PMC
parameters, whereas terdiurnal tidal structures are of minor importance.
Our analysis shows that the local time dependence becomes the most evident when
concentrating on one single season. When limiting the analysis to the season
2009 we find that local time variations of temperature at 69∘ N are in
a range of ±3 K near 83 km altitude. At sublimation altitudes (near
81.5 km) the water vapor variation is about ±3.5 ppmv. The variation in
water vapor leads to a change in the saturation ratio from about 1.8 around
midnight to 1 in the afternoon.
We calculated a climatology of IWC local time variations from a 35-year average
from 1979 to 2013 for different thresholds and latitude bands, which might be
useful for satellite data analysis in order to perform local time
corrections. Local time variations are found to depend on latitude and
threshold conditions. For the latitude band 64–74∘ N and a
threshold of IWC > 0 g km-2, IWC maximum and minimum values occur
around 03:00 and 19:00 LT, respectively, with a ratio of maximum to minimum of
6.6. For a threshold of IWC > 40 g km-2 the local times for maximum
and minimum are identical, but the ratio changes to 1.7. A phase shift exists
for the IWC local time behavior towards the pole, which is independent of the
threshold value. We find the absolute IWC local time variation generally
increases with latitude. Furthermore, the IWC maximum moves backward in time
from 08:00 LT at midlatitudes to 02:00 LT at high latitudes.
It should be noted that gravity waves could mask the influence of tides,
especially for the terdiurnal component. Gravity waves are partly included in
the MIMAS model, but a detailed investigation regarding their effects on the
tidal behavior of PMCs is beyond the scope of this paper. However, we expect
that the latitudinal variations of tidal amplitudes are robust and will help
in interpreting long-term observations with varying latitudes and fixed or
variable local times.
The model data are available upon request from the
corresponding author, and ALOMAR lidar data are available upon request from
Jens Fiedler. The AIM community provided access to the SOFIE version 1.3
downloaded from http://sofie.gatsinc.com/sofie/index.phpTand and CIPS Level 3c version 4.20, also freely available at
http://lasp.colorado.edu/aim/download-data-pmc.php.
Francie Schmidt drafted the paper. All the authors reviewed the paper and interpreted the data.
Uwe Berger conducted the LIMA and MIMAS modelling. Jens Fiedler had provided
the lidar data from ALOMAR.
The authors declare that they have no conflict of
interest.
This article is part of the special issue “Sources,
propagation, dissipation and impact of gravity waves (ACP/AMT inter-journal
SI)”. It is not associated with a conference.
Acknowledgements
We appreciate the financial support from the German BMBF for the ROMIC/TIMA
project. This research was supported by the European Union's Horizon 2020
Research and Innovation program under grant agreement no. 653980. The
European Centre for Medium-Range Weather Forecasts (ECMWF) is gratefully
acknowledged for providing ERA-40 and operational analysis
data. Edited by: Jörg
Gumbel Reviewed by: two anonymous referees
ReferencesBaumgarten, G., Fiedler, J., Lübken, F.-J., and von Cossart, G.: Particle
properties and water content of noctilucent clouds and their interannual
variation, J. Geophys. Res.-Atmos., 113, D06203,
10.1029/2007JD008884, 2008.Baumgarten, K., Gerding, M., Baumgarten, G., and Lübken, F.-J.: Temporal
variability of tidal and gravity waves during a record long 10-day continuous
lidar sounding, Atmos. Chem. Phys., 18, 371–384,
10.5194/acp-18-371-2018, 2018.Berger, U.: Modeling of middle atmosphere dynamics with LIMA, J.
Atmos. Sol.-Terr. Phy., 70, 1170–1200,
10.1016/j.jastp.2008.02.004, 2008.Berger, U. and Lübken, F.-J.: Trends in mesospheric ice layers in the
Northern Hemisphere during 1961–2013, J. Geophys. Res.-Atmos., 120, 11277–11298, 10.1002/2015JD023355, 2015.Berger, U. and von Zahn, U.: Icy particles in the summer mesopause region:
Three-dimensional modeling of their environment and two-dimensional modeling
of their transport, J. Geophys. Res., 107, 1366,
10.1029/2001JA000316, 2002.Berger, U. and von Zahn, U.: 3d-Modeling of the trajectories of visible NLC
particles indicates that these particles nucleate well below the mesopause,
J. Geophys. Res., 112, D16204, 10.1029/2006JD008106, 2007.Chu, X., Gardner, C. S., and Roble, R. G.: Lidar studies of interannual,
seasonal, and diurnal variations of polar mesospheric clouds at the South
Pole, J. Geophys. Res.-Atmos., 108, 8447,
10.1029/2002JD002524, 2003.Chu, X., Espy, P. J., Nott, G. J., Diettrich, J. C., and Gardner, C. S.:
Polar
mesospheric clouds observed by an iron Boltzmann lidar at Rothera (67.5S,
68.0W), Antarctica from 2002 to 2005: Properties and implications, J. Geophys. Res.-Atmos., 111, D20213, 10.1029/2006JD007086, 2006.DeLand, M. T. and Thomas, G. E.: Updated PMC trends derived from SBUV
data,
J. Geophys. Res.-Atmos., 120, 2140–2166,
10.1002/2014JD022253, 2015.DeLand, M. T., Shettle, E. P., Thomas, G. E., and Olivero, J. J.: Direct
observations of PMC local time variations by Aura OMI, J.
Atmos. Sol.-Terr. Phy., 73, 2049–2064,
10.1016/j.jastp.2010.11.019, 2011.
Douglass, A. and Fioletov, V.: Stratospheric ozone and surface ultraviolet
radiation, Global Ozone Research and Monitoring Project Report 52, chap. 2,
p. 76, World Meteorol. Organ., Geneva, Switzerland, 2011.Fiedler, J., Baumgarten, G., and von Cossart, G.: Noctilucent Clouds above
ALOMAR between 1997 and 2001: Occurrence and Properties, J.
Geophys. Res., 108, 8453, 10.1029/2002JD002419, 2003.Fiedler, J., Baumgarten, G., and von Cossart, G.: Mean diurnal variations of
noctilucent clouds during 7 years of lidar observations at ALOMAR, Ann.
Geophys., 23, 1175–1181, 10.5194/angeo-23-1175-2005, 2005.Fiedler, J., Baumgarten, G., and Lübken, F.-J.: NLC observations during
one solar cycle above ALOMAR, J. Atmos. Sol.-Terr. Phy., 71, 424–433,
10.1016/j.jastp.2008.11.010, 2009.Fiedler, J., Baumgarten, G., Berger, U., Hoffmann, P., Kaifler, N., and
Lübken, F.-J.: NLC and the background atmosphere above ALOMAR, Atmos.
Chem. Phys., 11, 5701–5717, 10.5194/acp-11-5701-2011, 2011.Fiedler, J., Baumgarten, G., Berger, U., and Lübken, F.-J.: Long-term
variations of noctilucent clouds at ALOMAR, J. Atmos. Sol.-Terr. Phy., 162, 79–89,
10.1016/j.jastp.2016.08.006, 2017.
Gadsden, M. and Schröder, W.: Noctilucent Clouds, vol. 18, edited by:
Lanzerotti, L. J., Hill, M. and Stöffler, D., Springer-Verlag, New York,
USA, 1998.Global Atmospheric Technologies and Sciences (GATS): AIM SOFIE data,
available at: http://sofie.gats-inc.com/sofie/index.php, last access:
25 November 2016.Gerding, M., Höffner, J., Lautenbach, J., Rauthe, M., and Lübken,
F.-J.: Seasonal variation of nocturnal temperatures between 1 and 105 km
altitude at 54∘ N observed by lidar, Atmos. Chem. Phys., 8,
7465–7482, 10.5194/acp-8-7465-2008, 2008.Gerding, M., Kopp, M., Hoffmann, P., Höffner, J., and Lübken, F.-J.:
Diurnal variations of midlatitude NLC parameters observed by
daylight-capable lidar and their relation to ambient parameters, Geophys.
Res. Lett., 40, 6390–6394, 10.1002/2013GL057955, 2013.Hartogh, P., Sonnemann, G. R., Song Li, Grygalashvyly, M., Berger, U., and
Lübken, F.-J.: Water vapor measurements at ALOMAR over a solar cycle
compared with model calculations by LIMA, J. Geophys. Res.-Atmos., 115,
D00I17, 10.1029/2009JD012364, 2010.Herbort, F., Baumgarten, G., Berger, U., Fiedler, J., Hoffmann, P., and
Lübken, F.-J.: Tidal structures within the LIMA model, Adv. Space Res.,
40, 802–808, 10.1016/j.asr.2007.04.061, 2007.Hervig, M. E. and Stevens, M. H.: Interpreting the 35 year SBUV PMC
record with SOFIE observations, J. Geophys. Res.-Atmos., 119, 12689–12705,
10.1002/2014JD021923, 2014.Hervig, M. E., Gordley, L. L., Stevens, M. H., III, J. M. R., Bailey, S. M.,
and Baumgarten, G.: Interpretation of SOFIE PMC measurements: Cloud
identification and derivation of mass density, particle shape, and particle
size, J. Atmos. Sol.-Terr. Phy., 71, 316–330,
10.1016/j.jastp.2008.07.009, 2009a.Hervig, M. E., Stevens, M. H., Gordley, L. L., Deaver, L. E., Russell, J. M.,
and Bailey, S. M.: Relationships between polar mesospheric clouds,
temperature, and water vapor from Solar Occultation for Ice Experiment
(SOFIE) observations, J. Geophys. Res.-Atmos., 114, D20203,
10.1029/2009JD012302,
2009b.Hervig, M. E., Siskind, D. E., Stevens, M. H., and Deaver, L. E.:
Inter-hemispheric comparison of PMCs and their environment from SOFIE
observations, J. Atmos. Sol.-Terr. Phy., 104,
285–298, 10.1016/j.jastp.2012.10.013, 2013.Hervig, M. E., Berger, U., and Siskind, D. E.: Decadal variability in PMCs
and implications for changing temperature and water vapor in the upper
mesosphere, J. Geophys. Res.-Atmos., 121, 2383–2392,
10.1002/2015JD024439,
2016a.Hervig, M. E., Gerding, M., Stevens, M. H., Stockwell, R., Bailey, S. M.,
Russell III, J. M., and Stober, G.: Mid-latitude mesospheric clouds and their
environment from SOFIE observations, J. Atmos. Sol.-Terr. Phy., 149, 1–14, 10.1016/j.jastp.2016.09.004,
2016b.Höffner, J. and Lübken, F.-J.: Potassium lidar temperatures and
densities in the mesopause region at Spitsbergen (78∘ N), J.
Geophys. Res., 112, D20114, 10.1029/2007JD008612, 2007.Jesse, O.: Die Höhe der leuchtenden Nachtwolken, Astron. Nachr., 140,
161–168, 10.1002/asna.18961401102, 1896.
Kiliani, J.: 3-D modeling of noctilucent cloud evolution and relationship to
the ambient atmosphere, PhD thesis, IAP Kühlungsborn, Kühlungsborn,
Germany, 2014.Kiliani, J., Baumgarten, G., Lübken, F.-J., Berger, U., and Hoffmann, P.:
Temporal and spatial characteristics of the formation of strong noctilucent
clouds, J. Atmos. Sol.-Terr. Phy., 104, 151–166, 10.1016/j.jastp.2013.01.005, 2013.Kiliani, J., Baumgarten, G., Lübken, F.-J., and Berger, U.: Impact of
particle shape on the morphology of noctilucent clouds, Atmos. Chem. Phys.,
15, 12897–12907, 10.5194/acp-15-12897-2015, 2015.Laboratory for Atmospheric and Space Physics (LASP): AIM CIPS data, available
at: http://lasp.colorado.edu/aim/download-data-L3C.php, last access:
5 December 2016.
Lindzen, R. and Chapman, S.: Atmospheric Tides, Space Sci. Rev., 10,
3–188, 1969.Lübken, F.-J.: Thermal structure of the Arctic summer mesosphere, J.
Geophys. Res.-Atmo., 104, 9135–9149, 10.1029/1999JD900076, 1999.
Lübken, F.-J., Fricke, K. H., and Langer, M.: Noctilucent clouds and the
thermal structure near the Arctic mesopause in summer, J. Geophys. Res., 101,
167–196, 1996.Lübken, F.-J., Baumgarten, G., Fiedler, J., Gerding, M., Höffner, J., and
Berger, U.: Seasonal and latitudinal variation of noctilucent cloud
altitudes, Geophys. Res. Lett., 35, L06801, 10.1029/2007GL032281, 2008.Lübken, F.-J., Berger, U., and Baumgarten, G.: Temperature trends in the
midlatitude summer mesosphere, J. Geophys. Res.-Atmos.,
118, 13347–13360, 10.1002/2013JD020576, 2013.Lumpe, J., Bailey, S., Carstens, J., Randall, C., Rusch, D., Thomas, G.,
Nielsen, K., Jeppesen, C., McClintock, W., Merkel, A., Riesberg, L.,
Templeman, B., Baumgarten, G., and Russell III, J. M.: Retrieval of polar mesospheric
cloud properties from CIPS: Algorithm description, error analysis and cloud
detection sensitivity, J. Atmos. Sol.-Terr. Phy.,
104, 167–196, 10.1016/j.jastp.2013.06.007, 2013.Megner, L.: Minimal impact of condensation nuclei characteristics on
observable Mesospheric ice properties, J. Atmos. Sol.-Terr. Phy., 73,
2184–2191, 10.1016/j.jastp.2010.08.006, 2011.Russell, J. M., Bailey, S. M., Gordley, L. L., Rusch, D. W., Horányi, M.,
Hervig, M. E., Thomas, G. E., Randall, C. E., Siskind, D. E., Stevens, M. H.,
Summers, M. E., Taylor, M. J., Englert, C. R., Espy, P. J., McClintock,
W. E., and Merkel, A. W.: The Aeronomy of Ice in the Mesosphere (AIM)
mission: Overview and early science results, J. Atmos. Sol.-Terr. Phy., 71, 289–299,
10.1016/j.jastp.2008.08.011, 2009.Russell, J. M., Rong, P., Hervig, M. E., Siskind, D. E., Stevens, M. H.,
Bailey, S. M., and Gumbel, J.: Analysis of northern midlatitude noctilucent
cloud occurrences using satellite data and modeling, J. Geophys.
Res.-Atmos., 119, 3238–3250, 10.1002/2013JD021017, 2014.Schöch, A., Baumgarten, G., and Fiedler, J.: Polar middle atmosphere
temperature climatology from Rayleigh lidar measurements at ALOMAR
(69∘ N), Ann. Geophys., 26, 1681–1698,
10.5194/angeo-26-1681-2008, 2008.Singer, W., Bremer, J., Hocking, W., Weiss, J., Latteck, R., and Zecha, M.:
Temperature and wind tides around the summer mesopause at middle and arctic
latitudes, Adv. Space Res., 31, 2055–2060,
10.1016/S0273-1177(03)00228-X, 2003.Singer, W., Latteck, R., Hoffman, P., Williams, B., Fritts, D., Murayama, Y.,
and Sakanoi, K.: Tides near the Arctic summer mesopause during the
MaCWAVE/MIDAS summer program, Geophys. Res. Lett., 32, L07S90,
10.1029/2004GL021607, 2005.Stevens, M. H., Englert, C. R., DeLand, M. T., and Hervig, M.: The polar
mesospheric cloud mass in the Arctic summer, J. Geophys. Res.-Space, 110,
A02306, 10.1029/2004JA010566, 2005.Stevens, M. H., Siskind, D. E., Eckermann, S. D., Coy, L., McCormack, J. P.,
Englert, C. R., Hoppel, K. W., Nielsen, K., Kochenash, A. J., Hervig, M. E.,
Randall, C. E., Lumpe, J., Bailey, S. M., Rapp, M., and Hoffmann, P.: Tidally
induced variations of polar mesospheric cloud altitudes and ice water content
using a data assimilation system, J. Geophys. Res.-Atmos., 115, D18209, 10.1029/2009JD013225, 2010.Stevens, M. H., Lieberman, R. S., McCormack, D. E. S. J. P., Hervig, M. E.,
and Englert, C. R.: Periodicities of polar mesospheric clouds inferred from a
meteorological analysis and forecast system, J. Geophys. Res.-Atmos., 122,
4508–4527, 10.1002/2016JD025349, 2017.
Thomas, G.: Is the polar mesosphere the miner's canary of global change?,
Adv. Space Res., 18, 149–158, 10.1016/0273-1177(95)00855-9, 1996.Thomas, G. E. and Olivero, J. J.: Climatology of polar mesospheric
clouds. II – Further analysis of Solar Mesosphere Explorer data, J.
Geophys. Res.-Atmos., 94, 14673–14681, 10.1029/JD094iD12p14673, 1989.Thomas, G. E., McPeters, R. D., and Jensen, E. J.: Satellite observations of
polar mesospheric clouds by the solar backscattered ultraviolet spectral
radiometer: Evidence of a solar cycle dependence, J. Geophys. Res.-Atmos.,
96, 927–939, 10.1029/90JD02312, 1991.Turco, R., Toon, O., Whitten, R., Keesee, R., and Hollenbach, D.: Noctilucent
clouds: Simulation studies of their genesis, properties and global
influences, Planet. Space Sci., 30, 1147–1181,
10.1016/0032-0633(82)90126-X, 1982.von Zahn, U. and Berger, U.: Persistent ice cloud in the midsummer upper
mesosphere at high latitudes: Three-dimensional modeling and cloud
interactions with ambient water vapor, J. Geophys. Res.-Atmos., 108, 8451,
10.1029/2002JD002409, 2003.
von Zahn, U., von Cossart, G., Fiedler, J., and Rees, D.: Tidal
variations
of noctilucent clouds measured at 69N latitude by groundbased lidar,
Geophys. Res. Lett., 25, 1289–1292, 1998.Witt, G.: Height, structure and displacements of noctilucent clouds, Tellus,
14, 1–18, 10.1111/j.2153-3490.1962.tb00115.x, 1962.