Slant total electron content (STEC) data measured by the Global Positioning System receiver at Ilorin, Nigeria, with geographical coordinates 8.47°N, 4.68°E for the year 2009 (a low-activity year) was used to study the diurnal, monthly standard deviation and monthly median value of total electron content (TEC). The vertical total electron content (VTEC) values are estimated from the STEC data. The thin shell approximation with an ionospheric shell height of 350 km was used for the analysis. The diurnal variation of VTEC (DTEC) and its corresponding monthly median variation (MTEC) shows a minimum at presunrise between the hours of 05:00 and 06:00 LT. The DTEC values show a maximum variation range from ∼24 to ∼34 total electron content unit (TECU). The daytime maximum TEC values observed in all the months were broad with a slight daytime depression in May, June, July, and November. The maximum variation of MTEC after slight daytime depression is greater than its variation before the slight daytime depression in the months affected with the month of July as exception. The slight daytime depression was lowest in the month of May and has a value of 0.99 TECU. A postsunset decrease at 20:00 LT with corresponding enhancement 2 h later was observed in the month of March. This post sunset decrease and enhancement in the month of March could be a strong indicator of the abrupt onset of scintillations, plasma bubbles, and spread F phenomenon. The monthly standard deviation depicts summary behavior of all the diurnal variations in each month. Annual and seasonal variations were also investigated.
 Global Positioning System (GPS) was originally designed in 1973 by the United States of America Department of Defense (DOD) for radio navigation [Misra and Enge, 2006]. Later, it gained wider use for relative positioning and timing. Accurate measurements for navigation, relative positioning, and time transfer could not be achieved at that time. This is because the radio signals while traversing through the ionosphere experienced delay coupled with satellite, receiver, and interchannel biases. With a careful choice of the elevation angle to aid better visibility of the GPS receiver to the satellites in space, multipath, which can also contribute to the inaccuracy of these measurements, could be avoided. These ideas were demonstrated in the work of Klobuchar et al.  when they used the single-frequency band of GPS and estimated the total electron content (TEC) from a model, which is marred with uncertainty effect on the final navigation solution. The ionospheric time delay at the L1 band (f1) as given by Klobuchar  is
where C is the speed of light in free space.
 In order to overcome this uncertainty in the final navigation solution, the single-frequency band (L1) which was then an important parameter for estimating TEC has been further augmented to dual-frequency bands (L1) and (L2) by the DOD. One of the advantage of the dual-frequency usage over a single-frequency receiver is that it operates on the 1.57542 GHz (L1) and 1.22760 GHz (L2) frequencies simultaneously. This operation permits measurement of the relative phase delay between the two signals, defined as the slant TEC, or the total number of electrons in a column of cross-sectional area of 1 m2 along the signal path between the satellite and the receiver. To the first order, it compensates for the ionospheric delay. GPS radio signals encounter time delay in the L1 and L2 simultaneous transmissions from GPS satellites through the ionosphere to the ground station. The dispersive nature of the ionosphere refracts radio signals, causing time delay [Ginzberg, 1970; Chen, 1984; Davies, 1989; Hofmann-Wellenhof et al., 2001]. The degree of refraction is proportional to the number of free electrons encountered by the signals, which is quantified as TEC [Rishbeth and Garriott, 1969]. The time delay measurements on the two L bands made it possible to compute the slant total electron content (STEC) from GPS data. These measurements of the true value of TEC were estimated to account for range delay and ionospheric correction in order to compute accurate position, velocity, and time. To the first order, the change in satellite-to-receiver signal propagation time due to the ionosphere is directly proportional to the integrated free electron density along the signal path. For the dual-frequency band, the difference in time delay between the two frequencies, Δt = t2 − t1, is given by
Thus, the time delay (Δt) measured between the L1 and L2 frequencies is used to calculate the STEC along the raypath. The pseudorange data provide absolute values of TEC but are noisy while the differential carrier phase gives a precise measure of the relative TEC variations, but the actual number of cycles of phase is not known. The idea about the combination of both advantages and disadvantages of pseudorange and pseudophase measurements resulted in absolute scale of TEC values and increases measurement accuracy.
 Before the era of GPS for TEC measurements, several researchers made significant contributions by estimating TEC with the use of orbiting satellites, such as BE-B, BE-C, Transit-4A, Transit-4B, and the geostationary satellite ATS-6 at low-latitude regions. In Africa, Skinner  and Olatunji  used the satellite signals of BE-B, Transit-4A, and Transit-4B to study the variations of TEC in Nigeria at Zaria (11.21°N, 7.71°E) and Ibadan (7.51°N, 3.91°E) using the data of 1964–1965 and 1962–1963, respectively. Olatunji  observed maximum value of TEC around 14:30 LT with no bite-out, while Skinner  observed two-peaked TEC maximum values during the daytime. In South America, Ross  analyzed the satellite signals of Transit-4A to study the TEC variation at Huancayo (11.21°S, 75.41°W), Peru, from 1961 to 1963 and found semiannual variation during daytime. In Asia, Rastogi et al.  calculated TEC at Thumba (8.51°N, 77.1°E), India, from the satellite signals of BE-B and BE-C for the period covering 1965–1968. They found that TEC value reaches a minimum around 05:00 LT and a broad maximum during daytime between 14:00 and 18:00 LT. Klobuchar and Rastogi  estimated TEC at Ootacamund (10.51°N, 72.51°E), India, from the signals of ATS-6 satellite between 1975 and 1976 while Rufenach et al.  studied TEC over Bangkok (13.71°N, 100.11°E), Thailand with the data of 1964 from the signal of Transit-4A. They both found the noontime maxima in the semithickness and equivalent slab thickness.
 As pointed out here, a number of authors who studied TEC at low-latitude sector have been presented. In this paper, we investigated the variability of GPS TEC at Ilorin, Nigeria, an equatorial zone, during the period of low solar activity and compared the results with previous studies.
2. Data and Method of Analysis
 Joint collaboration between the Ionospheric Physics and Radio Propagation Group, Department of Physics, University of Ilorin, Nigeria, and the Institute for Scientific Research, Boston College, Chestnut Hill, Massachusetts, United States, has brought about the installation of a GPS receiver at the University of Ilorin campus in 2008. The aim of the collaboration is to study the features of the Nigerian ionosphere in detail by using the continuous measurements of the TEC data collected at Ilorin (8.47°N, 4.68°E), Nigeria.
 The GPS data provide an efficient way to estimate TEC values with greater spatial and temporal coverage [Davies and Hartmann, 1997]. The GPS receiver can track up to 11 GPS coarse acquisition code signals at the L1 frequency of 1.575 GHz.
 The estimation of STEC by the above method in equation (2) was adopted by the GPS receiver. The estimated STEC data, S4 index, and other useful information like navigation and receiver positioning files are recorded at 1 min intervals. These automatic estimates and records of data are possible because of the software installed on the local computer, which is a collection of script written in Perl. In this paper, our focus is on variability of TEC data during low solar activity period; hence, STEC data recorded during 2009 has been employed.
 The STEC records from GPS are polluted with satellite differential delay, bS (satellite bias), and receiver differential delay, bR (receiver bias), coupled with receiver interchannel bias (bRX). This uncorrected STEC measured at every 1 min interval from the GPS receiver derived from all the visible satellites at the Ilorin station are converted to vertical TEC (VTEC). VTEC can be expressed as
where STEC is the uncorrected slant TEC measured by the receiver, S(E) is the obliquity factor with zenith angle, z, at the ionospheric pierce point (IPP), E is the elevation angle of the satellites in degrees, and VTEC is the vertical TEC at the IPP. The S(E) is defined by Mannucci et al.  and Langley et al.  as follows:
RE is the mean radius of the Earth measured in km and hS is the height of the ionosphere from the surface of the Earth, which is approximately equal to 350 km.
 These analyses from equations (3) and (4) were implemented in the C programming language developed by S. G. Krishna (Global Positioning System total electron content analysis application user's manual, 2009, Institute for Scientific Research, Boston College, Chestnut Hill, Massachusetts). This software reads raw data, processes cycle slips in phase data, reads satellite biases from International GNSS Service (IGS) code file (if not available, it calculates them), calculates receiver bias, and calculates the interchannel biases for different satellites in the receiver. To eliminate effect due to multipath, a minimum elevation angle of 20° is used. The VTEC data estimated are then subjected to a two sigma (2σ) iteration, which are a measure of GPS point positioning accuracy (95% confidence level). This 95% confidence level corresponds to 1.96 standard deviations, which is then approximate to two standard deviations or 2σ, and the resulting values are the average of VTEC over all pseudorandom numbers (PRNs) on a day. A sample of the software results for 1 November 2009 is shown in Figure 1.
 This software is used to compute the entire GPS STEC data for 2009 under investigation. These average 1 min VTEC data were converted to hourly values. The time convention for these analyses is in universal time (UT), but local time (LT) will be engaged in this paper. Nigeria is 1 h ahead of Greenwich mean time; therefore, 12:00 UT is 13:00 LT in Nigeria.
 The monthly median values (MTEC) and standard deviation (SD) values are deduced from DTEC values for their median and standard deviation values for each month. The annual variation of DTEC values are plotted from the first day of January to the last day of December 2009 on all the hours. The seasonal variation is grouped into four seasons: December solstice or D season (November, December, and January), March equinox or March E season (February, March, and April), June solstice or J season (May, June, and July), and September equinox or September E season (August, September, and October). Each season is estimated by averaging the MTEC values variations under a particular season. From the World Data Center catalog (http://wdc.kugi.kyoto-u.ac.jp/), on looking at the disturbance storm time (Dst) and planetary magnetic (Kp) indices throughout 2009, we found that this year was a solar quiet year and no major storm was produced except one moderate storm observed on 22 July 2009 (Dst = −78 nT and Kp = 4.3).
3. Results and Discussion
 The VTEC derived from the measured GPS STEC (as described in section 2) has been computed for all the days during 2009 at the Ilorin station. The number of days of data that has been used in each month is shown with the histogram graph in Figure 2. Time series line plots of the diurnal variations of DTEC for all months from the year 2009 along with their corresponding standard deviations and monthly median variations are presented in Figure 3. Figure 3 shows the DTEC during all days and MTEC values on the left-hand side of the y axis. The SD is plotted on the right-hand side of the y axis, and all values are then plotted together with local time on the x axis. The DTEC variations are given by green lines, MTEC variations are given by dashed black lines with asterisks, and SD variations are given by red lines with inverted triangles. There were a number of data gaps as a result of interruptions in the data recording, mostly due to power interruption. From Figure 2, April had about 6 days of data. Seventeen and eighteen days of data are available in February and May, respectively. For June and October, 24 days of data are plotted, and the remaining months are above 25 days of data. The contour diagram showing annual variation of DTEC for all the months on all hours of the day in 2009 at Ilorin is presented in Figure 4, and Figure 5 shows the seasonal variation. As can be observed in Figure 3, variability of DTEC, SD, and MTEC are characterized by minimum presunrise magnitudes, maximum daytime magnitudes, daytime depression, and postsunset depletion and enhancement.
3.1. DTEC, SD, and MTEC Variations
 In all the graphs in Figure 3, DTEC exhibits consistent minimum diurnal variation during presunrise hours between 05:00 and 06:00 LT, rises steeply during the sunrise period (07:00–09:00 LT), and subsequently rises very slowly from 10:00 LT to the peak during the daytime, mostly around 12:00–16:00 LT.
 The MTEC minimum value is in the range of ∼4–5 TECU between presunrise times of 05:00 and 06:00 LT. Maximum value of MTEC variation during the presunrise hour of 05:00 LT was observed in the month of May and has ∼5 TECU, while the minimum values of MTEC at the same presunrise hour were seen in February with ∼4 TECU.
Figure 3 reveals that the hour at which DTEC reaches the diurnal peak varies from day to day, and its daytime magnitude is greater than its postsunset and presunrise magnitudes. From all the months, the maximum DTEC daytime magnitude range is between ∼24 and ∼34 TECU, while the postsunset to presunrise magnitude range is between ∼1 and ∼7 TECU. The summary of the diurnal variations from day to day were seen in the SD variation of all of the months. From our study, the SD of TEC was found to closely respond to all the diurnal variability of TEC. The standard deviations are within the range of 0.1–3.5 TECU. They experience as many increases and decreases in all months in response to day-to-day variability of TEC as can be observed in Figure 3. These SD variabilities are observed throughout the period (01:00–24:00 LT) in all the months and follow the same pattern as the corresponding DTEC. They both showed sunrise minimum, daytime maximum, slight daytime depression, and postsunset decrease and enhancement. The exception to these was observed in April where there is a nighttime enhancement at 21:00 LT. This abnormal behavior of SD in April at nighttime could be a result of inadequate data during the month as can be observed in Figure 2. SD revealed clearer nighttime enhancement in February, March, September, October, November, and December, which can be seen in March only around 22:00 LT by looking at DTEC and MTEC variations. These SD nighttime enhancements of TEC ranged from 2.1 to 3.5 TECU around 20:00–22:00 LT.
 The daytime MTEC maximum values range is broad in the sense that it occurs between 13:00 and 16:00 LT with the exception of months that were affected as a result of slight depression. The months that were affected by slight daytime depression are May, June, July, and November. Therefore, MTEC maximum daytime magnitude range is between ∼21 TECU in August and ∼30 TECU in October. It is clearer that the range in maximums during daytime VTEC in all the months is greater than the range in minimums observed at nighttime (postsunset and presunrise). This greater magnitude of VTEC during daytime is attributed to solar EUV ionization coupled with the upward vertical E × B drift. These results are in close agreement with the results obtained at the equatorial region by Skinner , Olatunji , Rama Rao et al. , Liu and Chen , Liu et al. [2009a], Rastogi et al. , Davies and Hartmann , and Mendillo et al. . They concluded that the upward drift velocity can lift the ionosphere to higher altitudes where the ionization loss rate is smaller.
 As mentioned earlier, slight daytime depression characteristics observed in DTEC variation exist in the months of May, June, July, and November. The percentage ratio of the number of days slight depressions occur to the total number of the days in these months showed statistically that 79.70% of their daily TEC variations showed slight depression phenomenon during the daytime. These slight daytime depressions have different magnitudes and occurred at different times for all the days of the aforementioned months.
 The broad daytime maximum values of MTEC from 13:00–16:00 LT do not apply to May, June, July, and November because they exhibit slight daytime depression. This slight daytime depression occurred at 13:00 LT in June, 14:00 LT in May and July, and 15:00 LT in November. In May, the value of the daytime depression is 0.99 TECU. In June, July, and November, it decreases in value by 0.22, 0.45, and 0.59 TECU, respectively. These results showed that the daytime depression was highest in the month of May and lowest in the month of June. These results further revealed that the daytime postnoon peaks after these slight depressions are greater than the first peaks before the slight depressions, with an exception in the month of July.
 The results discussed above correspond to previous studies on prenoon and afternoon peak by Skinner , Bandyopadhyay , Rajaram , Lee and Reinisch , and Anderson and Klobuchar . They observed this similar phenomenon of daytime depression. They attributed this phenomenon to chemical loss of the daytime fountain effect at the magnetic equator. They further suggested that in hastening this chemical loss during daytime, regular chemical decay is augmented by downward driven motions of both poleward and zonal (eastward) neutral wind blowing away the fountain effect from the equator at noon until afternoon.
 The diurnal variability observed in March after sunset exhibits steep and sharp decrease of DTEC values and subsequent enhancement at later hours. This DTEC decrease was rapid in the month of March. The TEC dusk time variations in all months, with the exception of months when daytime depression occurs, are characterized by steep decrease in values until sunset hours (17:00–19:00 LT). The exception to this was observed in March, which is characterized by steep and sharp dusk time decrease until sunset. This DTEC decrease is followed by subsequent enhancement 2 h later. As mentioned earlier, these dusk time enhancements are clearer looking at the SD variation in each month. During the sunset hours, this steep decrease lasted for ∼1.5–3 h, after which postsunset MTEC values decayed gradually and smoothly on all the days of the month until presunrise hours.
 This signature does not conform to the postsunset characteristic that was observed in the month of March. In this case, the steep decrease is for about ∼1.5–2 h. In the month of March, sharp and rapid decrease in MTEC value was observed at 20:00 LT with subsequent enhancement at 22:00 LT. After the occurrence of postsunset enhancement at 22:00 LT, MTEC further decayed gradually and smoothly through the midnight until presunrise hours. This is evidence that the magnitude of MTEC postsunset variation is always greater than its presunrise variation.
 The contour plot in Figure 4 shows that the annual variation and DTEC maximum occurs between 13:00 and 18:00 LT. The unfilled (white background) contours in Figure 4 are a result of data gaps. From nighttime until presunrise hours (19:00–06:00 LT), minimum DTEC values are observed for all hours of the entire months. These minimum DTEC values range between ∼3 and ∼10 TECU. The highest maximum values of DTEC at peak period vary from ∼32 to ∼38 TECU around 14:00–15:00 LT. The highest maximum DTEC value was seen in October (∼38 TECU), followed by September (∼33 TECU), and was lowest in March (∼32 TECU). Other months also have high DTEC values at peak period but not as much as those mentioned. The color bar beside the contour plot in Figure 4 shows the level of their differences in magnitudes. Semiannual variation of DTEC values was observed in Figure 4; they are two peaks of maximum DTEC values. These two peaks occurred around March and October. Our results reasonably correspond to the previous works by Ross , Olatunji , Scherliess and Fejer , Bailey et al. , and Liu et al. [2007, 2009a, 2009b]. They observed semiannual variation in their work. Ross  suggested that the variation of the intensity of the equatorial electrojet obtained from the H component observation is mainly responsible for the semiannual DTEC variation. Scherliess and Fejer  suggested that daytime E × B drift velocities are larger in the equinoctial months (February, March, April, August, September, and October) and winter months (November, December, and January) than in the summer months (May, June, and July) and this could result in semiannual variation. Olatunji , Bailey et al. , and Liu et al. [2007, 2009a, 2009b] found that this semiannual variation is related to the variation of the noon solar zenith angle, which is an important factor for the production of ionization. Wu et al. , Rama Rao et al. , and Lee et al.  attributed the semiannual variation to a combined effect of solar zenith angle and magnetic field geometry.
3.3. Seasonal Variation
 In Figure 5, the seasonal variation depicts a highest value of 24.94 TECU in September equinox at 15:00 LT, followed by March equinox having 24.88 TECU at 14:00 LT. The solstice magnitudes are lower compared to the equinoxes, but the December solstice magnitude (24.08 TECU) is higher than the June solstice magnitude (21.03 TECU). Thus, the seasonal variation shows a semiannual pattern, maxima in equinoctial months and minima in solstice months. The possible mechanisms responsible for this pattern have been previously discussed in section 3.2. Also, DTEC maximum variation is observed earlier, between 14:00 and 16:00 LT in solstice months and later, between 15:00 and 16:00 LT in equinoctial months. These earlier and later DTEC maximum variations affirmed the sharpness in morning rise and afternoon decay of DTEC during the solstices compared to the equinoctial months. Between the solstices, the forenoon rate of production and afternoon decay of ionization is faster in the December solstice compared to that in the June solstice. The December solstice value is therefore greater than the June solstice. The DTEC minimum value of 3.20–7.21 TECU was observed at 06:00 LT, regardless of the season.
 This paper presents the variability of total electron content (TEC) over the equatorial West African region during the low solar activity period of 2009 using measurements of slant total electron content from the GPS receiver located at Ilorin (8.47°N, 4.68°E), Nigeria. The diurnal and statistical time series plots of VTEC shown in Figure 3 demonstrate that during the low solar activity period of 2009, the minimum value of MTEC, observed during presunrise periods (01:00–06:00 LT), ranged from 4.08 to 5.26 TECU, while the maximum value of DTEC, observed during the daytime period (13:00–16:00 LT), ranged from 23.67 to 33.88 TECU. SD depicts a summary measure of the day-to-day variation of TEC in each month. It also gave clearer representation of nighttime enhancements.
 Estimated DTEC and MTEC values gave clearer representation of the West African equatorial low-latitude features. Minimum values of DTEC and MTEC are frequent between 05:00 and 06:00 LT. Slight daytime depressions and postsunset decreases and enhancements were also observed. The slight daytime depression was highest in the month of May with a value of 0.99 TECU. The maximum daytime values after the slight depression in the affected months are greater than the maximum daytime values before the slight depression with the month of July as the exception.
 We would like to thank G. Krishna for helpful discussions on the GPS C-codes software used for this work. The donation of GPS facilities by the Boston College to the University of Ilorin, Nigeria, and the support of the University of Ilorin in running the GPS station are gratefully acknowledged. B.O.S. thanks the International Centre for Theoretical Physics (ICTP), Trieste, Italy, for granting him a Sandwiched Training Educational Programme (STEP) award toward his Ph.D. and the Aeronomy and Radio Propagation Laboratory (ARPL), where this work was successfully carried out.