The first Resolute incoherent scatter radar observations of the polar ionospheric F region show the fine control of the ionospheric plasma density and flow (electric field) by the solar wind interplanetary magnetic field (IMF). A summary of 8 days of observations is presented and 10 IMF Bz southward turning events during this period are analyzed in terms of the time delay of plasma density enhancements and ionospheric convection intensification with respect to the timing of Bz southward turning. We find that Ne enhancements are strongly tied to strong (≳5 nT) IMF Bz southward turnings; arrive 25–75 mins (depending on MLT) after the IMF pulse arrives at the bowshock nose; last as long as Bz stays southward; contain as small as ∼25 km horizontal substructures; are altitudinally smooth, a characteristic of a solar produced plasma. The most predictable response of ionospheric convection is anti-sunward flow intensification on average ∼25 mins after Bz southward change.
 SRI International through a cooperative agreement with the National Science Foundation has recently (in 2009) completed the construction of the northward-looking face of the Resolute incoherent scatter radar (RISR-N) in Resolute Bay, Canada. The new instrument can provide, for the first time, horizontal and altitudinal profiles of electron density to kilometric resolution, revealing the fine structure of polar plasma density irregularities (patches) as they drift across the electronically steered radar beams. Moreover, line-of-sight ion drift measurements at each beam position enable an analysis of the spatio-temporal variations. Most importantly, spacecraft measurements of solar wind parameters can be compared to RISR ionospheric measurements for prolonged durations to elucidate the detailed response of polar cap convection to solar magnetic field disturbances.
 In this letter we present the initial measurements for an 8-day period near equinox in the context of solar wind IMF parameters. First, we provide the RISR parameters and the methods used in extracting the electron density and ionospheric flow. Second, we compare the ionospheric measurements to solar wind parameters measured by the WIND spacecraft. We then select a set of Bz southward turning events and determine the time delays for Ne enhancements and flow intensifications. Finally, an individual plasma density enhancement lasting an hour is further resolved spatially and temporally.
2. Experiment Description
Table 1 shows the system parameters of RISR for this experiment. A total of 11 radar beams were used with the az/el coverage shown in Figure 1. Electron density profiles are obtained by averaging individual beam Ne measurements binned in altitude. The Ne measurements have been calibrated using independent Ne measurements using the plasma line technique. Mean ion drifts were obtained by least-squares fitting a uniform horizontal flow velocity vector to the observed line-of-sight velocities measured between the altitudes of 200–400 km.
Table 1. Resolute North Radar Parameters
Magnetic dip angle
Magnetic local time
Solar local time
Figure 2 shows the Ne altitude profile (averaged over all beam directions) as a function of UT day in September 2009. Superimposed on the color plot is the total electron content in TEC units (1 TEC = 1016 elec/m2) over the altitudes 100 to 700 km, the solar zenith angle (magenta line), and the IMF Bz in arbitrary scale. The background Ne follows a diurnal pattern with the maximum densities observed near local noon. On top of the background, Ne structures at ∼1 hour scales near the F peak are observed both day and night. The F peak density without (with) the Ne structures is ∼2.0 × 1011 m−3 (∼3.0 × 1011 m−3). A pattern of higher TEC or Ne structure occurrence during increased Bz magnitude is distinguishable.
Figure 3a (top) shows the IMF Bz (red) and By (blue) components measured by the WIND spacecraft on 2009/09/15–16. The data are provided online at the NASA OmniWeb site. (Note that the online IMF data are already time-shifted (by solar wind speed) from the spacecraft to the bowshock nose. The time-shifting procedure, e.g., determination of the bowshock nose location, phase front normals, is also described at OmniWeb.) We further delayed the By and Bz data shown in Figure 3a by 30 min to visually match the sharp Bz southward turnings to the leading edges of the Ne enhancements. This delay is on the order of solar wind travel time along the Sun-Earth line between the bowshock nose and the RISR magnetic field line that connects to the solar wind magnetic field at the flanks of the magnetotail for Bz southward. During this period, the solar wind velocity was ∼360 km/s. Figure 3a (middle) shows the Ne altitude profile as a function of UT with the color scale shown at the bottom of the plot. The black line is the total electron content in TEC units. Figure 3a (bottom) shows the anti-sunward (red line) and the dusk-to-dawn (green line) ion drift estimates from RISR. Figures 3b and 3c show the same data from days 9/20 and 9/21, except that the IMF data has been delayed by 40 and 25 min, respectively.
 An overview of Figure 3 is as follows: (1) The F region electron density enhancements are confined to the altitude range 250–400 km, mostly located at the F peak (300 km) and above. The peak density (including the background density) is 3 × 1011 m−3. The total TEC enhancements are up to 100%. The Ne enhancements appear to occur in response to every Bz southward turning event independent of local time. (2) The delays of the Ne enhancements are variable but still Bz events can be visually matched to Ne events, which are marked by up/down arrow heads. The duration of Ne enhancements is determined by how long Bz stays southward, e.g., on 9/15 between 2100–2400 UT. (3) We find somewhat good correlation between IMF Bz and sunward ionospheric flow in Figures 3b and 3c; however, the most predictable response of ionospheric convection is anti-sunward flow intensification to a strong Bz southward change. There exist periods showing strong correlation between Bz and sunward flow and, simultaneously, between By and dusk-to-dawn flow (e.g., 2009/09/15 1600–2000 UT, not shown here), correlation that is as good as shown by Hosokawa et al. . However, considering all the data between 9/15–9/23, and excluding the strong Bz southward turning events, the relationship between IMF and polar cap convection over Resolute is not as predictable as we expected from a simple 2-cell or 4-cell convection pattern.
 To further analyze the relationship between Bz-Ne and Bz-ionospheric convection, we have selected 10 events from Figure 3. The event triggers are marked by E1–E10 in Figures 3a (top), 3b (top), and 3c (top). We selected the beginning edges of density enhancements with an upward arrow for each event in Figures 3a (middle), 3b (middle), and 3c (middle). Also, we marked the ionospheric flow enhancement events with an arrow in Figures 3a (bottom), 3b (bottom), and 3c (bottom).
Table 2 provides the event statistics as the following parameters from left to right: Event no, day of the month, UT (for Ne event), MLT, solar wind velocity (Vsw), intensity of Bz southward turnings in nT (dBz), duration of Bz southward turnings WBz, duration of Ne enhancement (WNe), the delay from Bz to Ne (Ne delay), the delay from Bz to ionospheric flow enhancement (Flow delay).
Table 2 columns for WBz and WNe show a clear correlation between the durations of Ne enhancements and southward Bz. This data is plotted in Figure 4 (left).
Figure 4 (right) shows a distribution of Ne delays in MLT. The consistent pattern of Ne enhancements around noon arriving earlier than the ones at midnight points to the cusp origin (around 75° geomagnetic latitude) of the patches. Resolute Bay can be as close as 8° (900 km) in the local noon sector to as far as 22° (2400 km) in the midnight sector to the cusp. Assuming 500 m/s anti-sunward flow speed, and a delay of ∼10 min from the bowshock to the ionosphere [Yu and Ridley, 2009], the delay times would be 40 min to 90 min. These are roughly in agreement with the values in Table 2.
 The last column in Table 2 shows the ionospheric flow enhancement delays, which, in contrast to the Ne delays, appear to be independent of MLT (more statistics is needed to measure MLT dependence). The flow enhancements arrive (10–60 mins) earlier than the Ne enhancements. The average delay for the 10 events is 23 mins. Using the delay estimate of ∼10 min from the bowshock to the ionosphere by Yu and Ridley , we obtain an average delay of 13 mins, which falls in the range of mean delays between −3 and 17 mins by Khan and Cowley  based on the EISCAT data.
Figure 5 (top) shows a time-lapse (∼75 s) sequence of the altitude profile of an individual Ne enhancement corresponding to the event 8. The enhancement peak lies between 250–300 km altitude. The Ne profile shows no clear structure in altitude (fine altitude structure is expected in response to mono-energetic particle precipitation). Figure 5 (bottom) shows the time patterns of Ne(300 km) measured by the radar beams 7, 3, and 8 (see the beam locations in Figure 5). Note the sharp and rapid rise at beam 7 at minute 209 marked by the dashed vertical line. This is followed by rapid rises at beam 8 at minute 213 and at beam 3 at minute 220. The 11 min time delay between the first peaks in B7 and B3, which are horizontally separated by ∼300 km, implies a velocity of ∼400 m/s, which is in good agreement with the mostly eastward EXB drift velocity of ∼400 m/s observed during the event 8. The beam 3 pattern is less like the others. We think this is because the elevation of 35° is very low and, as a result, the pulse length of 480 μs, which is about 70 km horizontally, smears out the sharp horizontal gradients that may be as short as 20 km. This is not the case for beams 7 and 8 because of their relatively high (60°) elevation angles.
 We can infer the elongation of the Ne structure perpendicular to its motion. The first Ne(300 km) rise at beam 7 takes ∼1 min. For a ∼400 m/s flow, this means a scale of 25 km. Note that beams 7 and 8 are latitudinally separated by ∼150 km. The fact that the same rise is happening at beam 8 after proper eastward EXB delay means that the structure is elongated (>150 km) in the direction perpendicular to its motion, consistent with the prediction of patch formation theory by Lockwood and Carlson .
4. Discussion and Conclusion
 Some of the present findings are well in line with the existing knowledge on patches [Crowley, 1996]. It is well-known that plasma patches are observed in the polar cap F region occurring during Bz southward. The plasma patches are thought to originate near the cusp and drift (as a result of solar wind driven ionospheric convection) in the anti-sunward direction across the polar cap. The new RISR observations contribute to previous studies made by ionosondes, radio-tomography, and airglow imagers [e.g., MacDougall and Jayachandran, 2007; Walker et al., 1999; Hosokawa et al., 2009] by simultaneously resolving the altitudinal and horizontal structuring within the Ne enhancements to ∼25 km resolution. In summary, we find that
 1. Ne enhancements are strongly tied to strong (≳5 nT) IMF Bz southward turnings and last as long as Bz stays southward (0.5–2h in our observations). In this regard, polar patches can be regarded as tracers of strong solar wind “magnetic patches” crossing the Earth's magnetosphere.
 2. There exist substructures embedded within larger scale patches. For a select event, we observed ∼1 EXB minutes or 20–30 km horizontal structuring in the EXB direction within an Ne enhancements that lasted 90 mins. The sharp structures were found to be elongated at least ∼150 km in the direction of E.
 3. The pattern of time-delay from Bz southward turning moments to the Ne enhancements above RISR point to the cusp as their origin.
 4. Altitude profiles of the leading edge of an Ne enhancement event show no clear localization in altitude that is characteristic of particle precipitation; but rather, the enhanced altitude profiles are characteristic of a solar produced plasma.
 5. The most predictable feature of ionospheric convection is anti-sunward flow in response to strong Bz southward turns, compatible with the formation of the two-cell convection pattern. On average, the flow is not as simple as the two- or four-cell convection pattern.
 Our observations lend support to Lockwood and Carlson's  original hypothesis of polar cap Ne enhancements due to transient magneto-pause reconnection that occurs quasi-continuously every 1–2 minutes and that brings in solar produced plasma (as opposed to impact ionization due to cusp precipitation) across the polar cap. Lockwood and Carlson's  hypothesis has now been verified over the cusp in at least the European sector [Carlson et al., 2004]. Furthermore, Carlson et al.  documented the direct capture of co-rotating subauroral plasma to become injected into the polar cap; RISR observations of structures of similar nature show that their endurance over transport scales of 1000 km or more. Combining our observations with those on polar cap boundaries lends confirmation and deeper insight into the polar cap structuring processes intrinsic in transpolar plasma transport.
 Future studies will compare unique RISR observations of plasma density structures (provided by very rapid steering) to other radio wave (e.g., PolarDARN and Sondrestrom ISR) and optical measurements.
 RISR was developed under the NSF cooperative agreement ATM-0121483 to SRI International and the operations and maintenance is supported by NSF cooperative agreement ATM-0608577 to SRI International.