Geophysical Research Letters

Model simulations of global change in the ionosphere



[1] Observations of secular trends in the E and F1 regions of the ionosphere indicate that electron densities have increased, and that the height of the E-region peak has decreased, during the past several decades. Detection of trends in the upper ionosphere through analysis of F2-layer parameters has been more complex and controversial. In order to facilitate observational detection of long-term trends in the ionosphere, simulations were performed using a single-column upper atmosphere model. CO2 concentrations for the year 2000 and projected for the year 2100 were used to investigate changes of electron densities and the altitudes of ionospheric layers. Results show that increased CO2 concentration increases electron density in the lower regions of the ionosphere, but decreases electron density in the upper ionosphere. The transition altitude occurs slightly below the F2 peak altitude (hmF2). The proximity of hmF2 to the transition altitude may explain why different analyses of long-term trends in F2 peak density have shown both positive and negative trends. The altitudes of the E, F1 and F2 regions all decrease with increased CO2 concentration.

1. Introduction

[2] Roble and Dickinson [1989] first suggested that there would be long-term changes of the ionospheric E- and F-region peak densities in response to cooling and contraction of the mesosphere and the thermosphere, due to changes in greenhouse gas concentrations. Rishbeth [1990] concluded that long-term cooling in the upper atmosphere would lower the E- and F2-region peak heights but that changes in the E- and F2-region electron density should be small. Rishbeth and Roble [1992] investigated global change in the ionosphere using the NCAR Thermosphere-Ionosphere General Circulation Model (TIGCM) [Roble et al., 1988] by doubling CO2 and CH4 concentrations, and found that the height of the F2-layer peak dropped on average by about 15 km but that the F2-layer electron density change was minimal. Following these theoretical and modeling studies, ground-based ionosonde data have been analyzed in attempts to detect long-term trends in the ionosphere [e.g., Bremer, 1992; Ulich and Turunen, 1997; Bremer, 1998; Upadhyay and Mahajan, 1998; Danilov and Mikhailov, 1999; Mikhailov and Marin, 2000, 2001; Clilverd et al., 2003; Danilov, 2003]. There is evidence that the altitudes of the E and F1 regions have decreased, and that the E- and F1-region electron densities have increased in the past three to four decades [Bremer, 1998, 2001; Laštovička and Bremer, 2004; Laštovička et al., 2006a]. Detection of long-term trends of F2 parameters has been more complex and controversial with regard to methodology, results, and interpretation [Bremer, 1998; Mikhailov and Marin, 2000, 2001; Danilov, 2001; Clilverd et al., 2003; Danilov, 2003; Laštovička, 2005; Laštovička et al., 2006a, 2006b]. Both negative and positive trends of F2 peak electron density have been inferred; long-term changes of geomagnetic activity and greenhouse gas concentrations have been invoked as explanations.

[3] Causes of long-term trends of the ionosphere can be natural or anthropogenic in origin [Rishbeth, 1997]. Long-term changes in solar activity and geomagnetic activity could cause secular change in the ionosphere. Trends in the neutral atmosphere [Keating et al., 2000; Emmert et al., 2004; Marcos et al., 2005], which have been mainly attributed to increased concentration of CO2, can cause changes in the ionosphere through production, loss, and transport of the plasma. In addition, F2 region of the ionosphere is strongly influenced by dynamics and electrodynamics. Neutral winds or electric fields can shift the F2 peak altitude and change its density, and thus can cause complex geographic and solar local time distributions of any long-term trends.

[4] The purpose of this study is to explore mechanisms of trends in the ionosphere, to explain the complex pattern and controversies found in observations of trends of F2 parameters, and to facilitate ionosonde data analysis of long-term trends. Although secular trends in the past several decades may be influenced by natural causes, observed trends in the upper atmosphere and ionosphere have been mainly attributed to the steady increase of greenhouse gas concentrations [Danilov, 2003; Laštovička, 2005]. Therefore, this study focuses on investigation of ionospheric trends due to increased concentration of the greenhouse gas CO2. Long-term trends in the E and lower F1 regions may be slightly affected by stratosphere ozone depletion since the effect of ozone depletion on lower thermospheric temperature and density has been demonstrated by model study [Akmaev et al., 2006]. However, this secondary effect is not treated here. This paper addresses long-term trends of the global mean ionosphere rather than global distributions of the trends, and physical processes in the ionosphere are invoked to explain the changes in different ionospheric regions.

2. Model Description

[5] The model used in these studies is a self-consistent global mean model of the mesosphere, the thermosphere, and the ionosphere. Its original version is described by Roble et al. [1987]; extension to the mesosphere and additional modifications are documented by Roble [1995]. The model is a 1D representation of aeronomic processes in the NCAR Thermosphere-Ionosphere-Mesosphere-Electrodynamic General Circulation Model (TIME-GCM) [Roble and Ridley, 1994], and can be considered the single-column version of the TIME-GCM. Recent improvements include incorporation of a new solar EUV energy deposition scheme [Solomon and Qian, 2005], and updates to cooling rates and odd-nitrogen chemistry [Roble and Solomon, 2005].

[6] For present-day simulations, the model was run to steady state with a CO2 concentration of 365 ppmv imposed at the lower boundary (30 km), representative of the year 2000. For a future projection, a CO2 concentration of 730 ppmv was applied at the lower boundary, representative of the year 2100, as projected by the Intergovernmental Panel on Climate Change (IPCC) [2007] emission scenario A1B, a medium emission scenario. Different solar activity levels were employed but low geomagnetic activity conditions were assumed. Electron density profiles obtained by the different model experiments were then compared to quantify changes of ionospheric parameters. Mechanisms of these changes were explored, and solar-cycle dependence of these changes was also investigated.

3. Results

3.1. Long-term Trends in the E, F1, and F2 Regions

[7] In response to cooling and contraction of the middle atmosphere and thermosphere, the entire ionosphere will similarly contract, leading to lower altitudes of the E-region and F-region density peaks hmE and hmF2. Changes in the magnitudes of the density peaks NmE and NmF2, and changes in density throughout the ionosphere, are more subtle. Model experiments assuming moderate levels of solar activity (F10.7 = 150) were conducted to illustrate these changes.

[8] Figure 1a shows electron density profiles with CO2 concentrations at year 2000 level and the projected year 2100 level. The altitudes of the E and F1 regions decrease, which is consistent with the cooling and contraction of the thermosphere, while their electron density increases. Increase of electron density in the E- and F1-region electron density is controlled by compositional changes and subsequent changes in photochemical equilibrium. Section 3.2 provides a detailed discussion of the mechanisms. This pattern of decrease of altitude and increase of electron density in the lower ionosphere agrees with results from ionosonde data analyses [Bremer, 1998, 2001; Laštovička and Bremer, 2004; Laštovička et al., 2006a].

Figure 1.

(a) Electron number density profiles for the base case and the doubled CO2 case, under solar medium conditions (F10.7 = equation image10.7 = 150). Solid line, base case; dotted line, doubled CO2 case. (b) Percentage change of electron number density from the doubled CO2 case to the base case, under solar medium conditions (F10.7 = equation image10.7 = 150).

[9] The F2 region contains both positive and negative change regions, but the peak electron density NmF2 decreases with increased CO2 forcing. The transition altitude from positive change to negative change is near but slightly below the peak altitude hmF2. This altitude indicates a transition from control by photochemical equilibrium to plasma transport, as discussed in section 3.2. Changes in forcing of photochemical equilibrium or forcing of transport can potentially cause a shift in this altitude. The proximity of the positive-to-negative transition to hmF2 may explain why ionosonde data analyses have resulted in a complex pattern of long-term trends in F2 peak electron density, with both positive and negative trends detected. Recent reviews of long-term trends of F2 peak electron density by Laštovička et al. [2006b, 2008] concluded that long-term trends of NmF2 are either negative or insignificant. This pattern of long-term trends of NmF2 is consistent with these model simulations. Figure 1a also shows that the altitude of the F2 region, including hmF2, decreases with increased CO2 concentration.

3.2. Mechanisms of Global Change in the Ionosphere

[10] The model experiments described in section 3.1 were analyzed to elucidate the reasons for the changes shown in Figure 1. The increased CO2 concentration causes electron density increases in the lower ionosphere up to about 250 km; above that altitude, electron density decreases. To understand the causes of this transition, simplified physics of the quasi-equilibrium ionosphere are described below.

[11] Electron density in the E region is proportional to ionization rate but inversely proportional to the ratio of the two major ions in the E region, NO+ and O2+. The E region is in approximate photochemical equilibrium between photoionization and loss of ions through dissociative recombination:

equation image

where M2 represents major species in the E region, which are O2 and N2, q is the ionization rate, and α is the effective dissociative recombination rate. Since N2+ is rapidly converted to NO+ and O2+ through reactions with atomic and molecular oxygen, the major ions in the E region are NO+ and O2+. Assuming photochemical equilibrium, q = α[M2+][e], where [M2+] and [e] are ion and electron number density, respectively. Assuming charge neutrality, [e] = equation image1/2. Since the dissociative recombination rate of NO+ is nearly twice that of O2+, the effective recombination rate α is proportional to NO+/O2+, and electron density is inversely proportional to this ratio:

equation image

[12] Figure 2b shows the corresponding altitude profile of percentage changes of the ionization rate and the ratio of NO+ and O2+ from year 2000 to year 2100, between 100 km and 130 km. Change of the ratio of NO+ to O2+ is negative in the E region. This is because lower temperature due to increased CO2 concentration causes lower NO density, which in turn decreases NO+ density since the main source of NO+ is the reaction of O2+ with NO. Possible negative secular trends of NO+/O2+ in the E region have been observed [Danilov, 1997, 2001]. Change of the ionization rate is positive at all E-region altitudes. Cooling and contraction of the thermosphere causes less absorption of solar irradiance before it reaches E-region altitudes, and thus enhances ionization rates in this region. Consequently, electron density in this altitude range increases with increasing CO2 concentration.

Figure 2.

(a) Percentage change of O/N2 from the base case to the doubled CO2 case in the altitude range 120 km to 250 km, under solar medium conditions (F10.7 = equation image10.7 = 150). (b) Percentage changes from the base case to the doubled CO2 case for the total ionization rate and the ratio of NO+ number density to O2+ number density in the E region, under solar medium conditions (F10.7 = equation image10.7 = 150).

[13] As the ionosphere transitions from the E region to the F region, atomic oxygen becomes the major neutral species. The F region up to the F2 peak is also in approximate photochemical equilibrium, but different photochemical production and loss processes dominate. The production of ions is mostly due to ionization of atomic oxygen. Since the recombination rate of O+ is slow, loss of ions goes through two steps: transfer of O+ to NO+ and O2+ through atom-ion interchange reactions followed by dissociative recombination of the molecular ions. The atom-ion interchange reactions are much slower than the dissociative recombination rates. Consequently, electron density is determined by the balance between the ionization rate and the rates of the atom-ion interchange reactions. Since N2 dominates over O2, electron density is roughly proportional to the ratio of O and N2, i.e., [e] ∝ O/N2. Figure 2a shows altitude profile of percentage change of O/N2 from 2000 to 2100 scenario in this region. Change of O/N2 is positive (on constant altitude surfaces) due to the cooling and contraction of the atmosphere, which yields an increase in electron density.

[14] The F2 peak occurs where transport of the plasma becomes comparable to chemical production and loss. Above the F2 peak, the electron density altitude profile becomes mainly controlled by plasma transport. Consequently, electron density decreases exponentially with the plasma scale height. Since the cooling and thus contraction of the atmosphere reduces the plasma scale height, electron density decreases responding to increased CO2 concentration in this altitude range.

3.3. Solar Cycle Dependence of Ionospheric Changes

[15] Thermospheric long-term trends show a clear solar-cycle dependence, with the secular decrease of neutral density under solar minimum conditions approximately three times of that under solar maximum conditions [Emmert et al., 2004; Qian et al., 2006]. It is reasonable to expect that the ionospheric long-term trends will also depend on solar activity. This was investigated using model experiments with CO2 concentrations for 2000 and 2100 as projected by the IPCC [2007] A1B emission scenario, under solar minimum and solar maximum conditions (F10.7 = 70, 200). Table 1 shows changes of E- and F2-region parameters as a result of increased CO2 concentration from 2000 to 2100. hmF2 decreases by 14 km and 10 km under solar minimum and solar maximum conditions, respectively, while NmF2 decreases by 9% and 4%. Therefore, trends of F2 parameters are expected to be larger under solar minimum conditions, similar to the solar-cycle dependence of secular change in the thermosphere. hmE decreases 2 km under solar minimum conditions and 4 km under solar maximum conditions. Changes of NmE are relatively small, with 4% and 2% increases under solar minimum and solar maximum, respectively.

Table 1. Changes of Ionospheric Parameters Due to Change of CO2 Concentration From 2000 to the Projected 2100 Concentration Based on the IPCC Emission Scenario A1B, Under Solar Minimum and Solar Maximum Conditions
 Solar MinSolar Max
hmF2−14 km−10 km
hmE−2 km−4 km

4. Discussion

[16] Model experiments show that long-term trends of electron density transitions from positive at lower altitudes to negative at higher altitudes responding to increased CO2 concentrations; the transition altitude is near but below the F2 peak. The positive and negative trends are determined by basic ionospheric physics, with positive trends in altitude ranges governed by photochemical equilibrium, and negative trends at higher altitude where plasma transport dominates.

[17] The E and F1 regions are within the positive change range and thus electron density in these regions should increase with increasing CO2 concentration. The E-region peak altitude, however, should decrease with increased CO2 concentration. This pattern of increase of electron density and decrease of the altitude of the lower ionosphere is consistent with results from analyses of ionosonde data attempting to measure long-term trends in the E and F1 regions during the past several decades. The F2 region consists of both positive and negative change regions, with the transition altitude near the F2 peak altitude. This may explain the complex results obtained from different analyses of ionosonde observations of secular change in F2 peak electron density, with both positive and negative trends found. Model simulation of the F2 peak electron density predicts that it decreases with increased CO2 concentration, in agreement with recent reviews of ionosonde trend detection. These simulations show that the F2 peak altitude also decreases with increased CO2 concentration.

[18] Trends in ionospheric F2 parameters are larger under solar minimum conditions than solar maximum conditions, similar to the solar-cycle dependence of secular trends in the thermosphere. This can be explained by examining the effects of stronger cooling under solar minimum conditions. The larger temperature decrease due to doubling of CO2 under solar minimum conditions causes a larger decrease in plasma scale height, which in turn causes a fixed pressure surface to descend more compared to the case of solar maximum conditions. Since hmF2 tends to remain on the same pressure surface as temperature changes [Garriott and Rishbeth, 1963], hmF2 will have a larger decrease under solar minimum conditions. In addition, the F2 peak is where chemical control of plasma density gives way to transport control. Plasma transport causes plasma density to decrease according to plasma scale height. Therefore, a larger change in plasma scale height under solar minimum conditions can result in greater change in NmF2.

[19] These results may give some guidance to further analyses of long-term ionospheric measurements. Changes in hmE may be a significant fraction of a scale height in the E-region, but since scale heights are low at these altitudes, this may be difficult to measure on decadal periods. Changes in NmE are small but detectable [Bremer, 2001; Bremer et al., 2004]. The secular trends in NmF2 may be comparatively small compared to intrinsic variability in the F-region, but this parameter is the most unambiguous and stable measurement that can be obtained from the extensive ionosonde historical data set, and is worth tracking carefully. The accurate measurement of hmF2 may be more difficult using ionosonde measurements, but it is obtainable from incoherent scatter radar observations, and it is the most direct and significant secular change that is expected from increasing CO2 levels.

[20] The upper ionosphere is highly influenced by neutral atmosphere dynamics and by electrodynamics. These effects could cause significant regional features in F2 long-term trends. For instance, long-term changes in the geomagnetic field could significantly influence secular change in the ionosphere in some geographic regions. Just as simulations of climate in the troposphere and at the Earth's surface are progressing from general averages of expected changes to regional climate change, studies of the effects in the thermosphere/ionosphere system must similarly progress. Three-dimensional modeling studies may improve our interpretation and understanding of secular trends inferred at particular geographic locations.


[21] This research was supported by NASA grants NNX07AC55G, NNX07AC61G, and NNH05AB55I to the National Center for Atmospheric Research. NCAR is supported by the National Science Foundation.