Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
 In the low-latitude, ionospheric F region, the primary transport mechanism that determines the electron and ion density distributions is the magnitude of the daytime, upward E × B drift velocity. During large geomagnetic storms, penetration of high-latitude electric fields to low latitudes can often produce daytime, vertical E × B drift velocities in excess of 50 m/s. Employing a recently developed technique, we can infer these daytime, upward E × B drift velocities from ground-based magnetometer observations at Jicamarca and Piura, Peru, as a function of local time (0700–1700 LT). We study the ionospheric response in the Peruvian longitude sector to these large upward drifts by theoretically calculating electron and ion densities as a function of altitude, latitude, and local time using the time-dependent Low-Latitude Ionospheric Sector (LLIONS) model. This is a single-sector ionosphere model capable of incorporating data-determined drivers, such as E × B drift velocities. For this study, we choose three large storms in 2003 (29 and 30 October and 20 November) when daytime E × B drift velocities approached or exceeded 50 m/s. Initial results indicate that the large, upward E × B drift velocities on 29 October produced equatorial anomaly crests in ionization at ±20° dip latitude rather than the usual ±16° dip latitude. We compare the theoretically calculated results with a variety of ground-based and satellite observations for these three periods and discuss the implications of these comparisons as they relate to the capabilities of current theoretical models and our ability to infer ionospheric drivers such as E × B drifts (Anderson et al., 2002).
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The daytime equatorial electrojet is a narrow band of enhanced eastward current flowing in the 100–120 km altitude region within ±2° latitude of the dip equator. A unique way of determining the daytime strength of the electrojet is to observe the difference in the magnitudes of the horizontal (H) component between a magnetometer placed directly on the magnetic equator and one displaced 6°–9° away [Kane, 1973; Rastogi and Klobuchar, 1990]. The difference between these measured H values, ΔH, provides a direct measure of the daytime electrojet current and in turn, the magnitude of the vertical E × B drift velocity in the F region ionosphere.
 A recent paper [Anderson et al., 2004a], based on the Anderson et al.  study, discussed a study where 27 months of magnetometer H component observations and daytime, vertical E × B drift velocities were obtained in the Peruvian longitude sector between August 2001 and December 2003.
 The ΔH values correspond to the difference in the horizontal (H) component between magnetometers located at Jicamarca (11.9°S geographic latitude, 283.1°E geographic longitude, 0.8° dip latitude) and Piura (5.2°S geographic latitude, 279.4°E geographic longitude, 6.8° dip latitude), Peru. The daytime, vertical E × B drift velocities were inferred from the Jicamarca Unattended Long-term Ionosphere Atmosphere (JULIA) radar at Jicamarca, Peru [Hysell et al., 1997]. In order to establish the relationships between ΔH and E × B drift velocities for the 270 days of observations, three approaches were chosen: (1) a linear regression analysis, (2) a multiple-regression approach, and (3) a neural network approach. The neural network method gave slightly lower RMS error values compared with the other two methods. Since the exact nature of the nonlinearity between the inputs and the E × B drift velocity observations is not known, the neural network approach can decipher these in a way that the multiple-regression technique cannot [Anderson et al., 2004a]. The relationships for all three techniques were validated using an independent set of E × B drift observations from the Jicamarca Incoherent Scatter Radar (ISR) located at Jicamarca, Peru. Figure 1, taken from Anderson et al. [2004a], compares the three techniques with the ISR observations of daytime, vertical E × B drift velocities on 17 April 2002. For the three storm days studied in this paper, neither the JULIA nor the Jicamarca ISR E × B drift velocities were available; hence the ΔH technique is required.
2. Brief Description of the Low-Latitude Ionospheric Sector Model
 The Low-Latitude Ionospheric Sector (LLIONS) model is a single-sector model that solves the coupled ion momentum and continuity equations and theoretically calculates ion and electron densities as a function of altitude, latitude, and local time. Since LLIONS does not self-consistently couple the ionosphere, thermosphere, and electrodynamics, inputs to LLIONS include the Mass Spectrometer Incoherent Scatter (MSIS) neutral atmospheric densities and temperatures, the Hedin neutral wind velocities, and low-latitude drivers such as vertical E × B drift velocities as a function of local time (see section 5 below). LLIONS is identical to the Utah State University Model of the Global Ionosphere [Schunk and Sojka, 1996] except that it is applicable for a single longitude sector. The longitude is specified as an input to LLIONS and can be any value from 0° to 360° geographic longitude.
3. Brief Description of the Geomagnetic Storm Periods in 2003
 The three specific geomagnetic storm periods this paper focuses on are the Halloween storms of 29 and 30 October 2003 and the 20 November 2003 storm. The 3-hour Kp values for the period 29 through 31 October were greater than 6 for the entire period, except for three 3-hour segments when they were greater than 5. The daily ΣKp for 29 October was 189 and for 30 October was 162. Preceding these two days, an X17 X-ray flare, the largest to date of solar cycle 23, occurred on 28 October. On 20 November, the daily ΣKp was 117. A detailed report of all solar and geomagnetic indices can be found in a NOAA service assessment report (Intense space weather storms October 19–November 07, 2003, U.S. Department of Commerce, April 2004, available at http://www.sec.noaa.gov/AboutSEC/index.html; go to Program Assessments).
4. Magnetometer-Inferred E × B Drift Velocities
 Ground-based magnetometers at Jicamarca (0.8°N dip latitude) and Piura (6.8°N dip latitude), Peru, provided the ΔH values that were used to calculate the daytime, vertical E × B drift velocities between 0700 and 1700 LT. The trained neural network approach developed by Anderson et al. [2004a] and discussed in section 1 was used to infer the daytime E × B drift velocities for the 3 storm days. Figure 2 displays these drifts as a function of local time. It is assumed the E × B drifts are independent of latitude.
 For each of the 3 days, the green curve represents the Fejer-Scherliess quiet time, climatological, vertical E × B drift velocity as a function of local time [Scherliess and Fejer, 1999]. Between 0700 and 1700 LT, the blue curve is the ΔH-inferred E × B drift velocity for each of the 3 days. On 29 and 30 October, between 1700 and 2100 LT, the Jicamarca digital sounder was used to infer the vertical E × B drift velocity using a technique developed by Anderson et al. [2004b]. Basically, the time rate of change in the bottomside height of the 2 × 105 el/cm3 density contour (∼4 MHz frequency) gives the vertical E × B drift velocity. On 20 November, the Jicamarca sounder was not operating. The red circles represent the E × B drift velocity every 15 min using this technique, and the smaller circles are the interpolated drift velocities every 5 min.
 The LLIONS model was used to theoretically calculate the electron and ion densities as a function of altitude (90–1600 km), latitude (35°N to 35°S dip latitude), and local time (24 hours) in the Peruvian longitude sector (∼290°E geographic longitude). Input parameters included (1) neutral atmospheric densities and temperatures from MSIS90 [Hedin, 1991]; (2) meridional and zonal neutral wind velocities from the MSIS-Wind model [Hedin et al., 1991]; (3) appropriate production, loss, and diffusion rates and Kp and F10.7 cm flux values; and (4) the E × B drift velocities pictured in Figure 2. We calculate the total electron content (TEC) values as a function of dip latitude and local time by integrating the electron density profiles from 90 to 1600 km altitude, every 2° in latitude and every 15 min in local time over the 24-hour day. These calculated TEC values are then compared with TEC observations obtained from a network of ground-based, dual-frequency GPS receivers in the South American longitude sector. The inputs to the MSIS90 and the MSIS-Wind models were the appropriate Kp and F10.7 cm flux values during each of the 3-day periods.
Figure 3 compares contours of theoretically calculated TEC values as a function of dip latitude and local time for 29 and 30 October assuming (1) the climatological, Fejer-Scherliess vertical E × B drift velocities and (2) the E × B drift velocities pictured in Figure 2. For the results displayed in Figure 3, Figure 3a shows the LLIONS-calculated TEC values versus geomagnetic latitude and local time, incorporating the climatological Fejer-Scherliess model (green curve) for the 24-hour day on 29 October. Figure 3b displays the LLIONS-calculated TEC values for 29 October when the E × B drifts are assumed to be (1) depicted by the dark blue, ΔH-inferred values between 0700 and 1700 LT, (2) depicted by the circles between 1700 and 2100 LT, and (3) depicted by the climatological green curve between 2100 and 0700 LT. The same E × B drift inputs are incorporated into LLIONS for the “climatological” day on 30 October (Figure 3c) and the “disturbed” day drifts on 30 October (Figure 3d).
 There are several features to note.
 1. Comparing Figures 3a and 3b, on 29 October at 1800 LT, the TEC crests are located at ±20° dip latitude (Figure 3b), while they normally occur at ±15° dip latitude (Figure 3a). This is a direct consequence of the large upward E × B drift velocities between 1300 and 1700 LT shown in Figure 2a.
 2. The ΔH-inferred E × B drifts observed on 29 and 30 October were basically downward between 0700 and 1130 LT (Figures 3b and 3d). This is responsible for the relatively late development of the equatorial anomaly crests compared with the “climatology” days (Figures 3a and 3c). In addition, the downward drifts are also responsible for the lower calculated TEC values at 1300 LT near the magnetic equator (70 TEC units (1 TECU = 1016 el/m2)) versus 90 TEC units for the “climatology” days 29 and 30 October).
 3. The effects of the E × B drift, prereversal enhancement features (1700–2100 LT) displayed in Figures 2a and 2b are evident when comparing the LLIONS results at 2000 LT on 29 October (Figures 3a and 3b) and 30 October (Figures 3c and 3d). On both days, the prereversal enhancement for the climatological drift (light green curves) has a maximum upward value of 50 m/s at 1900 LT, while the E × B drifts inferred from the Jicamarca sounder (circles) is 50 m/s downward at 2000 LT on 29 October. The consequence of the strong upward climatological drift is to enhance the TEC values to 180 TEC units at 2000 LT as displayed in Figures 3a and 3c. The absence of strong prereversal enhancement on both 28 and 30 October displayed by the circles in Figure 2a and 2b results in the absence of a TEC enhancement between 2000 and 2100 LT shown in Figures 3b and 3d.
 In Figure 4, we compare the theoretically calculated contours of TEC as a function of dip latitude and local time with GPS observations of equivalent, vertical TEC values from a chain of GPS dual-frequency receivers in South America. The E × B drift inputs to LLIONS that have already been described also describe the inputs for Figures 3b and 3d.
 The agreement between the calculated and observed TEC for 29 October is fairly good. The nighttime values of 10–30 TEC units are in general agreement, as are the daytime values of 90 TEC units near the dip equator. Both of the peak values in TEC at the crests are about 180 TEC units, and the location of the crests at ∼±18°–20° dip latitude are similar. On 30 October, the agreement is not as good, with the calculated TEC values at the crests displaying a greater Southern Hemisphere crest, while the observed TEC values from the South American chain of GPS receivers show a pronounced enhancement of 180 TEC units in the northern crest and only 80 TEC units in the southern crest.
 In Figure 5, we compare the DMSP F13 Special Sensor for Ions, Electrons and Scintillation (SSIES) observations of electron density versus dip latitude at 1830 LT and 840 km altitude with LLIONS results for 29 and 30 October. In Figure 5, the green curve represents the LLIONS results when the climatological Fejer-Scherliess E × B drifts are incorporated (Ne_LLIONS), the blue curve represents the LLIONS results when the neural network ΔH-inferred E × B drifts are incorporated (Ne_NN), and the red curve is the DMSP observations.
 On 29 October, two crests in electron density on either side of the dip equator are clearly observed by DMSP F13 with a density of 1.6 × 106 el/cm3 in the Southern Hemisphere and a density of 1.1 × 106 el/cm3 in the Northern Hemisphere. The calculated electron densities corresponding to the observations are 1.1 × 106 el/cm3 in both hemispheres, but the crests are not as widely separated in dip latitude as the observations. On 30 October, two crests are again observed by DMSP with densities > 6 × 105 el/cm3, but the crests are much closer together at ±8° dip latitude rather than ±15° dip latitude observed on 29 October. In the calculated densities on 30 October, there is only one maximum at the magnetic equator with a density of 5.5 × 105 el/cm3. The implication of the comparisons on 29 and 30 October is that the upward E × B daytime drift velocities should be slightly greater to bring about an improved comparison. Note that on both 29 and 30 October, the LLIONS “climatological” drift case (green curve) has a peak density of ∼ 1 × 106 el/cm3 near the dip equator with only a slight indication of crest separation.
 In Figure 6 we present the LLIONS calculated TEC values as a function of dip latitude and local time and a comparison with the ground-based GPS/TEC observations from the GPS receivers for the 20 November 2003 geomagnetic storm incorporating the E × B drift velocities shown in Figure 2c. The large, upward E × B drift velocities on 20 November produce crests in TEC at ±20° dip latitude by 1600 LT with calculated TEC values greater than 180 TEC units. These are in rough agreement with observed TEC values obtained from the South American network of GPS receivers.
Figure 7 compares the calculated values of electron density (Ne) (840 km) versus dip latitude with the DMSP/SSIES observations at 1830 LT for 20 November. While the DMSP F13 observations of electron density versus dip latitude display crests on either side of the dip equator, the observed values of ∼4 × 10 5 el/cm3 are significantly lower than the calculated values of ∼1 × 106 el/cm3 in the Southern Hemisphere and ∼7 × 105 el/cm3 in the Northern Hemisphere.
Figure 8 graphically displays the chain of ground-based, dual-frequency GPS receivers in the South American longitude sector that were used to provide equivalent, vertical TEC values for comparison with the theoretically calculated values.
 The purpose of this study was to investigate the low-latitude ionospheric response to three large geomagnetic storms and to compare this response to both ground-based and satellite observations. The daytime response is critically dependent on the vertical E × B drift patterns for these storms. A new technique for inferring these vertical E × B drift velocities on the basis of ground-based magnetometer observations is briefly described, and the E × B drift patterns for the three geomagnetic storms are presented. Comparing theoretically calculated ionospheric parameters from the LLIONS model with ground-based GPS observations of TEC values and DMSP/SSIES in situ Ne values at 840 km demonstrates that when realistic values of vertical E × B drift values are incorporated, the comparisons are in reasonable agreement. However, achieving closer agreement with the ground-based and satellite observations will require better knowledge of (1) neutral composition, (2) neutral meridional wind velocities, and (3) ion and neutral temperature values than are provided by the climatological models.