Features of the 11-year variation of galactic cosmic rays in different periods of solar magnetic cycles

Variations of galactic cosmic ray intensity have been studied based on the neutron monitors and interplanetary magnetic field experimental data for different ascending and descending epochs of solar activity. The dependence of the diffusion coefficient on the cosmic ray particles rigidity R is stronger in the maxima epoch than in the minima epoch of solar activity. For the period of 1977–1981 (qA > 0) the diffusion coefficient for the minimum epoch is, χ_min ∝ R^{0.7 ± 0.04} and for the maximum χ_max ∝ R^{1.3 ± 0.05}; for the period of 1987–1990 (qA < 0), χ_min ∝ R^{0.8 ± 0.05} and χ_max ∝ R^{1.1 ± 0.04}. The exponents ν_y and ν_z of the power spectral density of the B_y and B_z components of the IMF in the region of the frequencies (10⁻⁶– 4 × 10⁻⁶) Hz are larger for the minimum epoch of 1987 (ν_y ≈ 2.0 and ν_z ≈ 1.93) than for the maximum epoch of 1990 (ν_y ≈ 1.43 and ν_z ≈ 1.27).     

Statistical parameters of nonisothermal lower ionospheric plasma in the electrically active mesosphere

The large V/m electric fields inherent in the lower mesosphere play an essential role in lower ionospheric electrodynamics. They must be the cause of large variations in the electron temperature and the electron collision frequency and consequently of the transition of the ionospheric plasma in the lower part of the D region into a nonisothermal state. This study is based on the datasets on large mesospheric electric fields collected with the 2.2-MHz radar of the Institute of Space and Atmospheric Studies, University of Saskatchewan, Canada (52°N geographic latitude, 60.4°N geomagnetic latitude), and with the 2.3-MHz radar of the Kharkiv V. Karazin National University, Ukraine (49.6°N geographic latitude, 45.6°N geomagnetic latitude). The statistical analysis of these data is presented by [Meek, C.E., Manson, A.H., Martynenko, S.I., Rozumenko, V.T., Tyrnov, O.F. Remote sensing of mesospheric electric fields using MF radars. J. Atmos. Solar-Terr. Phys. 66, 881–890, 2004. 10.1016/j.jastp.2004.02.002]. The large mesospheric electric fields in the 60–67-km altitude range are experimentally established to follow a Rayleigh distribution in the 0 < E < 2.5 V/m interval. These data have permitted the resulting differential distributions of relative disturbances in the electron temperature, θ, and the effective electron collision frequency, η, to be determined. The most probable θ and η values are found to be in the 1.4–2.2 interval, and hence the nonstationary state of the lower part of the D region needs to be accounted for in studying processes coupling the electrically active mesosphere and the lower ionospheric plasma.     

Spectral analysis of a two-ribbon microflare

Using high-resolution Hα, CaII 8542 Å and FeI 6302.5 Å Stokes spectral data obtained simultaneously with THEMIS in 2002 September, we have analyzed the spectra and the characteristics of a two-ribbon microflare (MF). The hard X-ray emission provides evidence of non-thermal particle acceleration in the microflare. The two-ribbons are located on either sides of the magnetic polarity inversion line. The non-thermal characteristics mainly appeared at the outer edges of the flare ribbons. It indicates that the instantaneous magnetic reconnection and the particle acceleration mainly took place at the outer edges of the flare ribbons. Using the Hα and CaII 8542 Å line profiles and the non-LTE calculation, we obtain the semi-empirical atmospheric model for the bright kernel of the MF. The result indicates that the temperature enhancement in the chromosphere is about 2000–2500 K.     

Satellite monitoring of anomalous effects in the ionosphere probably related to strong earthquakes

The paper presents simultaneously a satellite monitoring of plasma density disturbances and quasi-static electric fields on the one hand, and a ground earthquake monitoring, on the other. The INTERCOSMOS-BULGARIA-1300 satellite was launched on 7th August 1981. It had a perigee of 825 km, an apogee of 906 km and orbit inclination of 81.2°. For satellite’s orbits over sources of earthquakes M ⩾ 5.5 (seismic data of the World Data Center – Denver, Colorado), in the time period 12.08.1981–30.12.1981, some ionospheric effects are investigated. These effects can be considered eventually as possible earthquake precursors. The paper discusses specific anomalous effects observed in the ionosphere, which cannot be explained by factors of solar-magnetospheric origin. They could be associated with the earthquake growing processes in the lithosphere and troposphere zones above the source. In conclusion the results of ionosphere anomalous effects monitoring are proposed.     

On long-term changes of electron density in the upper mesosphere

Recent review study done jointly by 19 experts of 17 institutes shows zero trend of temperature in the upper mesosphere. In the light of this latest development, we have examined the long-term changes in electron density, [e], in this region. The study has been concentrated at 80 km. At this altitude, electrons are mainly produced by the interaction of nitric oxide, NO, by solar Ly-α. Any long-term change in this flux will affect trend of [e]. Considering this flux proportional to 10.7 cm solar flux, analysis of available 10.7 cm solar flux data from 1948 to 2003 has been made. A decreasing trend up to about 1970 and then an increasing trend are found. The over-all increasing trend of Ly-α flux during the past five decades is ∼0.17% per year. This increase also gives a ∼0.17% increasing trend per year in [e]. This non-anthropogenic increase is much less compared to the observed increase in [e] which is reported to be >0.7% per year. The observed increase in [e] of this magnitude will then, predominantly, be due to the anthropogenic effect. In zero trend in temperature, significant change in electron loss coefficient, α_eff, and [NO] are unlikely to take place to cause a significant change in [e]. The increase in [e] > 0.7% per year then can be explained by considering a decreasing trend in [O₂].     

Propagation of secondary antiprotons and cosmic rays in the Galaxy

Recent measurements of the cosmic ray (CR) antiproton flux have been shown to challenge existing CR propagation models. It was shown that the reacceleration models designed to match secondary to primary nuclei ratios (e.g., B/C) produce too few antiprotons. In the present paper, we discuss one possibility to overcome these difficulties. Using the measured antiproton flux and B/C ratio to fix the diffusion coefficient, we show that the spectra of primary nuclei as measured in the heliosphere may contain a fresh local “unprocessed” component at low energies perhaps associated with the Local Bubble, thus decreasing the measured secondary to primary nuclei ratio. The independent evidence for SN activity in the solar vicinity in the last few Myr supports this idea. The model reproduces antiprotons, B/C ratio, and elemental abundances up to Ni (Z ⩽ 28). Calculated isotopic distributions of Be and B are in perfect agreement with CR data. The abundances of three “radioactive clock” isotopes in CR, ¹⁰Be, ²⁶Al, ³⁶Cl, are all consistent and indicate a halo size z_h ∼ 4 kpc based on the most accurate data taken by the ACE spacecraft.     

F2 layer characteristics and electrojet strength over an equatorial station

The data presented in this work describes the diurnal and seasonal variation in hmF2, NmF2, and the electrojet current strength over an African equatorial station during a period of low solar activity. The F2 region horizontal magnetic element H revealed that the Solar quiet Sq(H) daily variation rises from early morning period to maximum around local noon and falls to lower values towards evening. The F2 ionospheric current responsible for the magnetic field variations is inferred to build up at the early morning hours, attaining maximum strength around 1200 LT. The Sq variation across the entire months was higher during the daytime than nighttime. This is ascribed to the variability of the ionospheric parameters like conductivity and winds structure in this region. Seasonal daytime electrojet (EEJ) current strength for June solstice, March and September equinoxes, respectively had peak values ranging within 27–35 nT (at 1400 LT) , 30–40 nT (at 1200 LT) and 35–45 nT (at 1500 LT). The different peak periods of the EEJ strength were attributed to the combined effects of the peak electron density and electric field. Lastly, the EEJ strength was observed to be higher during the equinoxes than the solstice period.     

Gamma-ray observations of explosive nucleosynthesis products

In this review I discuss the various γ-ray emission lines that can be expected and, in some cases have been observed, from radioactive explosive nucleosynthesis products. The most important γ-ray lines result from the decay chains of ⁵⁶Ni, ⁵⁷Ni, and ⁴⁴Ti. ⁵⁶Ni is the prime explosive nucleosynthesis product of Type Ia supernovae, and its decay determines to a large extent the Type Ia light curves. ⁵⁶Ni is also a product of core-collapse supernovae, and in fact, γ-ray line emission from its daughter product, ⁵⁶Co, has been detected from SN1987A by several instruments. The early occurrence of this emission was surprising and indicates that some fraction of ⁵⁶Ni, which is synthesized in the innermost supernova layers, must have mixed with the outermost supernova ejecta.Special attention is given to the γ-ray line emission of the decay chain of ⁴⁴Ti (⁴⁴Ti → ⁴⁴Sc → ⁴⁴Ca), which is accompanied by line emission at 68, 78, and 1157 keV. As the decay time of ⁴⁴Ti is ∼86 yr, one expects this line emission from young supernova remnants. Although the ⁴⁴Ti yield (typically 10⁻⁵–10⁻⁴M_⊙) is not very high, its production is very sensitive to the energetics and asymmetries of the supernova explosion, and to the mass cut, which defines the mass of the stellar remnant. This makes ⁴⁴Ti an ideal tool to study the inner layers of the supernova explosion. This is of particular interest in light of observational evidence for asymmetric supernova explosions.The γ-ray line emission from ⁴⁴Ti has so far only been detected from the supernova remnant Cas A. I discuss these detections, which were made by COMPTEL (the 1157 keV line) and BeppoSAX (the 68 and 78 keV lines), which, combined, give a flux of (2.6 ± 0.4 ± 0.5) × 10⁻⁵ ph cm⁻² s⁻¹ per line, suggesting a ⁴⁴Ti yield of (1.5 ± 1.0) × 10⁻⁴M_⊙. Moreover, I present some preliminary results of Cas A observations by INTEGRAL, which so far has yielded a 3σ detection of the 68 keV line with the ISGRI instrument with a flux that is consistent with the BeppoSAX detections. Future observations by INTEGRAL-ISGRI will be able to constrain the continuum flux above 90 keV, as the uncertainty about the continuum shape, is the main source of systematic error for the 68 and 78 keV line flux measurements. Moreover, with the INTEGRAL-SPI instrument it will be possible to measure or constrain the line broadening of the 1157 keV line. A preliminary analysis of the available data indicates that narrow line emission (i.e., Δv < 1000 km s⁻¹) can be almost excluded at the 2σ level, for an assumed line flux of 1.9 × 10⁻⁵ ph cm⁻² s⁻¹.     

A search for radio emission from the young 16-ms X-ray pulsar PSR J0537−6910

PSR J0537−6910 is a young, energetic, rotation-powered X-ray pulsar with a spin period of 16 ms located in the Large Magellanic Cloud. We have searched for previously undetected radio pulsations (both giant and standard) from this pulsar in a 12-h observation taken at 1400 MHz with the Parkes 64-m radio telescope. The very large value of the magnetic field at the light cylinder radius suggests that this pulsar might be emitting giant radio pulses like those seen in other pulsars with similar field strengths. No radio emission of either kind was detected from the pulsar, and we have established an upper limit of ∼25 mJy kpc² for the average 1400-MHz radio luminosity of PSR J0537−6910. The 5σ single-pulse detection threshold was ∼750 mJy for a single 80-μs sample. These limits are likely to be the best obtainable until searches with greatly improved sensitivity can be made with next-generation radio instruments.     

The spatial distribution of galactic and anomalous cosmic rays in the heliosphere at solar minimum

The spatial distributions of galactic and anomalous cosmic rays in the heliosphere at the solar minima of Cycles 20/22 (qA > 0) and of Cycle 21 (qA < 0) are studied, using data from IMP 8, Voyagers 1/2 and Pioneer 10. It is found that the radial dependences of intensities J can be approximated by a power of radial distance r as J ∝ r^α with a different value of a constant in the inner and outer heliosphere with a transition at a radial distance of 10–15 AU. To study the physical meaning of these radial intensity profiles we examined the rigidity dependences of the intensity gradients by determining the particle mean free paths, using a simple one-dimensional modulation model. The particle mean free path λ was assumed to be a separable function of distance of the form r^γ and rigidity R of R^δ over the range of 0.5–3.0 GV in the inner and outer heliosphere. It was shown that λ of rigidity dependence of R^1.6 determined for Cycle 20/22 (qA > 0) with anomalous He is about 4–5 times larger than that of Cycle 21 (qA < 0) with R^0.9 at around 1 GV in the outer heliosphere, and that the radial dependences are r^1.4 and r^1.1, respectively, for Cycles 20/22 and for Cycle 21.     

Hard X-ray footpoint motions in solar flares: Comparing magnetic reconnection models with observations

Hard X-ray observations from the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI) of the October 29, 2003 GOES X10 two-ribbon flare are used together with magnetic field observations from the Michelson Doppler Imager (MDI) onboard SoHO to compare footpoint motions with predictions from magnetic reconnection models. The temporal variations of the velocity v of the hard X-ray footpoint motions and the photospheric magnetic field strength B in footpoints are investigated. The underlying photospheric magnetic field strength is generally higher (B ∼ 700–1200 G) in the slower moving (v ∼ 20–50 km s⁻¹) western footpoint than in the faster (v ∼ 20–100 km s⁻¹) moving eastern source (∼100–600 G). Furthermore, a rough temporal correlation between the HXR flux and the product vB² is observed.     

Excitations of nonmigrating diurnal tides in the mesosphere and lower thermosphere simulated by the Kyushu-GCM

Excitation mechanisms of nonmigrating diurnal tides in the MLT region simulated by the Kyushu-GCM are examined. It is shown that the westward propagating diurnal tide with zonal wavenumber s = 2 is mainly excited by nonlinear interactions between the migrating diurnal tide and the stationary planetary wave with zonal wavenumber s = 1, while the nonlinear excitation of the standing diurnal tide with zonal wavenumber s = 0 is less important than the excitation by tropospheric heating. Nonlinear interactions between the migrating diurnal tide and the stationary planetary wave with zonal wavenumber s = 2 are not dominant to excite the westward propagating diurnal tide with zonal wavenumber s = 3, and it is shown that the excitation by tropospheric heating is comparable to the nonlinear excitation. It is also shown that other nonmigrating diurnal tides are excited by tropospheric heating.     

Scaling and universality of the thermal conductivity of liquid 4He near the superfluid transition and in restricted geometries

We present measurements of the thermal conductivity λ(t, P, L) = l/ρ(t, P, L) near the superfluid transition of ⁴He at saturated vapor pressure and confined in cylindrical geometries with radii L = 0.5 and 1.0 μm (t ≡ T/T_λ(P) − 1). For L = 1.0 μm measurements at six pressures P are presented. At and above T_λ the data are consistent with a universal scaling function F(X) = (L/ξ_o)^x/ν(ρ/ρ₀), X = (L/ξ_o)^1/νt valid for all P (ρ₀ and x are the pressure-dependent amplitude and effective exponent of the bulk resistivity ρ(t, P, ∞) = ρ₀t^x and ξ = ξ₀t^−ν is the correlation length). Indications of breakdown of scaling and universality are observed below T_λ.     

Hypervelocity impacts on HST solar arrays and the debris and meteoroids population

Accurate debris and meteoroid flux models are crucial for the design of manned and unmanned space missions. For the most abundant particle sizes smaller than a few millimetres, knowledge of the populations can only be gained from in situ detectors or the analysis of retrieved space hardware. The measurement of impact flux from exposed surfaces improves with increased surface area and exposure time.A post-flight impact investigation was initiated by the European Space Agency to record and analyse the impact fluxes and any potential resulting damage on the two flexible solar arrays of the Hubble Space Telescope. The arrays were deployed during the first Hubble Space Telescope servicing mission in December 1993 and retrieved in March 2002. They have a total exposed surface area of roughly 120 m², including 42 m² covered with solar cells. This new Hubble post-flight impact study follows a similar activity undertaken after the retrieval of one of the first solar arrays, in 1993. The earlier study provided the first opportunity for a numerical survey of damage to exposed surfaces from more than 600 km altitude, and of impacts from particles larger than 1 mm. The results have proven very valuable in validation of important flux model regimes. The second set of Hubble solar arrays has again provided an unrivalled opportunity to measure the meteoroid and debris environment, now sampled during a long interval in low Earth orbit, and to identify changes in the space debris environment since the previous survey. The retrieved solar array wings exhibit thousands of craters, many of which are visible to the naked eye. A few hundred impacts have completely penetrated the 0.7 mm thick array. The largest impact features are about 7–8 mm in diameter. The cover glass of the solar cells is particularly well suited to the recognition of small impact features by optical and electron microscopy. In this paper, we present the first results of the impact survey. Data upon the abundance of craters of specific measured size ranges are plotted as cumulative flux curves, and compared to the results of model predictions. The most significant change to the particle flux since 1993 is a decrease in the small debris population.     

Impacts on Hubble Space Telescope solar arrays: Discrimination between natural and man-made particles

A Post-Flight Investigation was initiated by the European Space Agency to analyze impacts on solar arrays of the Hubble Space Telescope (HST), exposed to space for 8.25 years at approximately 600 km altitude. The solar cells deployed during the first Service Mission (SM-1 in December 1993) were retrieved in March 2002 as part of Service Mission 3B (SM-3B). A sub-panel of 2 m² was cut from the arrays for subsequent selection and removal of individual solar cells for analysis. Six cells (4.8 × 10⁻³ m²) were surveyed for flux of all craters of sizes greater than 5 microns. Analytical scanning electron microscopy was used to analyse residues in 111 features of 3–4000 micron conchoidal detachment diameter (D_co), examined on 23 solar cells. Eighty three show identifiable residue: 38 are Space Debris impacts and 45 Micrometeoroid impacts. Of the remaining 28, 2 contain residue of ambiguous origin, 1 is probably a minor manufacturing flaw, 1 is obscured by contamination, and 24 are unresolved, lacking recognizable residue.The majority of space debris impacts on the SM-3B cells are less than 80 microns D_co, dominated by Al-rich residue, probably of solid rocket motor origin, although three may be due to sodium metal droplet impacts. Three larger features include paint pigment and binder, ferrous alloy, and possible carbon-fibre composite material debris.Micrometeoroid residues are found across the entire crater size range and dominate features of between 100 and 1000 microns, their residues are similar to those found in earlier SM-1 surveys. Fe- and Mg-rich silicates dominate; Fe sulphides are common and there are occasional vesicular Ni- and S-bearing mafic silicates of hydrous phyllosilicate origin. A single sodium aluminosilicate residue and one Fe Ni metal residue were found; as well as enigmatic Mg- and S-bearing residues, all considered as probably of micrometeoroid origin. A few Fe-, O- and C-bearing residues were classified as of ambiguous origin.     

The role of drift and convection–diffusion mechanisms in small energy global cosmic ray modulation: Application to the hysteresis phenomenon of proton and alpha particle satellite data

The hysteresis effect for small energies of galactic cosmic rays is due to two effects. The first is the same as for neutron monitor energies – the delay of the interplanetary processes responsible for cosmic ray modulation with respect to the initiating solar processes, according to the effective velocity of solar wind and shock waves propagation. Then, the observed cosmic ray intensity is connected to the solar activity variations during many months before the time of cosmic ray measurement. The second is caused by the time delay of small energy cosmic ray diffusion from the boundary of modulation region to the Earth’s orbit. The model describing the connection between solar activity variation and cosmic ray convection–diffusion global modulation for neutron monitor energies is here developed by taking into account also the time-lag of the small energy particle diffusion in the Heliosphere. We use theoretical results on drifts and analytically approximate the dependences of drifts from tilt angle, and take into account the dependence from the sign of primary particles, and from the sign of polar magnetic field (A > 0 or A < 0). The obtained results are applied on proton and alpha-particle satellite data. We analyze satellite 5-min data of proton fluxes with energies >1 MeV, >2 MeV, >5 MeV, >10 MeV, >30 MeV, >50 MeV, >60 MeV, >100 MeV, and in intervals 10–30 MeV, 30–60 MeV, and 60–100 MeV during January 1986–December 1999. We exclude periods with great cosmic ray increases caused by particle acceleration in solar flare events. Then, we determine monthly averaged fluxes, as well as 5-month and 11-month smoothed data. We analyze also satellite 5-min data on alpha-particle fluxes in the energy intervals 60-160 MeV, 160–260 MeV and 330–500 MeV during January 1986–May 2000. We correct observation data for drifts and then compare with what is expected according to the convection–diffusion mechanism. We assume different dimensions of the modulation region (by the time propagation X₀ of solar wind from the Sun to the boundary of modulation region), for X₀ values from 1 to 60 average months, by one-month steps. For each value of X₀ we determine the correlation coefficient between variations of expected and observed cosmic ray intensities (the estimation of cosmic ray intensities values is given in Section 3 by Eq. (9), and the determination of correlation and regression coefficients in Section 3 by Eq. (8)). The dimension of modulation region is determined by the value of X_0 max, for which the correlation coefficient reaches the maximum value. Then the effective radial diffusion coefficient and residual modulation in small energy region can be estimated.     

A Gemini observation of the anomalous X-ray pulsar 1RXS J170849-400910

The anomalous X-ray pulsars (AXPs) represent a growing class of neutron stars discovered at X-ray energies. While the nature of their multi-wavelength emission mechanism is still under debate, evidence has been recently accumulating in favor of their magnetar nature. Their study in the optical and infrared (IR) wavelengths has recently opened a new window to constrain the proposed models. We here present a brief overview of AXPs and our Gemini-South observation of 1RXS J170849-400910, which is a relatively bright AXP discovered with ROSAT and later found to be an 11 s X-ray pulsar by ASCA. The observation was taken with the near-IR imager Flamingos in J (1.25 μm), H (1.65 μm), and K_s (2.15 μm). We confirm the recent detection by (ApJ, 589, L93–L96) of a source coincident with the CHANDRA source (candidate ‘A’). Our derived magnitudes of J = 20.6 (0.2), H = 18.6 (0.2), and K_s = 17.1 (0.2) are consistent with those derived by (ApJ, 589, L93–L96), and indicate that if this source is indeed the IR counterpart to 1RXS J170849-400910, then there is no evidence of variability from this AXP. However, given the lack of IR variability and the relatively high IR to X-ray flux of this source when compared to the other AXPs, we conclude that this source is unlikely the counterpart of the AXP, and that the other source (candidate ‘B’) within the CHANDRA error circle should not be ruled out as the counterpart. Further monitoring of these sources and a deep observation of this complex field are needed to confirm the nature of these sources and their association with the AXP.     

XMM-Newton observations of the Mouse,SLX 1744–299 and SLX 1744–300

We observed the radio and X-ray source G359.23–0.82, also known as “the Mouse”, with XMM-Newton. The X-ray image of this object shows a point-like source at the Mouse’s “head”, accompanied by a “tail” that extends for about 40″ westward. The morphology is consistent with that observed recently with Chandra [Gaensler, B.M., van der Swaluw, E., Camilo, F., et al. The Mouse that soared: high resolution X-ray imaging of the pulsar-powered bow shock G359.23–0.82, ApJ 616, 383–402, 2004]. The spectrum of the head can be described by a power-law model with a photon index Γ ≃ 1.9. These results confirm that the Mouse is a bow-shock pulsar wind nebula (PWN) powered by PSR J1747–2958. We found that the hydrogen column density toward the Mouse, N_H = (2.60 ± 0.09) × 10²² cm⁻², is 20%–40% lower than those toward two serendipitously detected X-ray bursters, SLX 1744–299 and SLX 1744–300. At a plausible distance of 5 kpc, the X-ray luminosity of the Mouse, L(0.5–10 keV) = 3.7 × 10³⁴ erg s⁻¹, is 1.5% of the pulsar’s spin-down luminosity. We detected a Type I X-ray burst from SLX 1744–300 and found a possible decrease of N_H and persistent luminosity for this source, in comparison with those observed with ROSAT in 1992.     

Supernova remnant 1987A: The latest report from the Chandra X-ray Observatory

We continue monitoring supernova remnant (SNR) 1987A with the Chandra X-ray Observatory. As of 2004 January, bright X-ray spots in the northwest and the southwest are now evident in addition to the bright eastern ring. The overall X-ray spectrum, since 2002 December, can be described by a planar shock with an electron temperature of ∼2.1 keV. The soft X-ray flux is now 8 × 10⁻¹³ ergs cm⁻² s⁻¹, which is about five times higher than four years ago. This flux increase rate is consistent with our prediction based on an exponential density distribution along the radius of the SNR between the HII region and the inner ring. We still have no direct evidence of a central point source, and place an upper limit of L_X = 1.3 × 10³⁴ ergs s⁻¹ on the 3–10 keV band X-ray luminosity.     

Detection of long-term trends in the mesosphere/lower thermosphere from ground-based radio propagation measurements

Mesospheric temperature trends can be derived from LF phase-height observations in mid-latitudes supported by ionospheric absorption and ionosonde observations. Analysing the full observation period from 1959 until 2003, a mean yearly temperature trend has been derived with −0.25 K/yr for the height interval from 48 to 82 km. Subdividing the whole observation interval in two parts before and after 1979, the trend is markedly stronger in the second period with −0.38 K/yr compared with −0.20 K/yr in the first part before 1979. These differences can at least partly be explained by a steeper CO₂ increase and ozone decrease in the second interval. The differences in the mesospheric temperature trends are most evidently expressed during winter months and are markedly smaller during summer season. The reason of this seasonal difference is not quite clear; it may be related with detected ozone trends which are clearly stronger during winter months on both hemispheres.