Evolution of High‐Speed Jets and Plasmoids Downstream of the Quasi‐Perpendicular Bow Shock

Plasma structures with enhanced dynamic pressure, density, or speed are often observed in Earth's magnetosheath. We present a statistical study of these structures, known as jets and fast plasmoids, in the magnetosheath, downstream of both the quasi‐perpendicular and quasi‐parallel bow shocks. Using measurements from the four Magnetospheric Multiscale (MMS) spacecraft and OMNI solar wind data from 2015–2017, we present observations of jets during different upstream conditions and in the wide range of distances from the bow shock. Jets observed downstream of the quasi‐parallel bow shock are seen to propagate deeper and faster into the magnetosheath and on toward the magnetopause. We estimate the shape of the structures by treating the leading edge as a shock surface, and the result is that the jets are elongated in the direction of propagation but also that they expand more quickly in the perpendicular direction as they propagate through the magnetosheath.


Introduction
The bow shock is the boundary at which the speed of the solar wind abruptly drops as it approaches Earth. Downstream of the shock, in the magnetosheath, the plasma is denser and hotter than in the unperturbed solar wind, and the magnetic field is stronger than the interplanetary magnetic field (IMF). The structure of the bow shock depends to a large extent on whether the IMF is close to parallel or perpendicular to the shock normal. This also affects the particle populations through processes that lead to reflection of particles, the formation of a foreshock, and the wave activity in the vicinity of the shock. For small angles between the bow shock normal and the IMF (θ Bn < 45°), the shock is quasi-parallel, and for large angles (θ Bn > 45°), it is known as quasi-perpendicular.
Ions reflected at the quasi-parallel shock travel as beams through the upstream plasma, generating waves in the foreshock region through wave-particle interaction (e.g., Wilson et al., 2013). This is also where short large-amplitude magnetic structures (SLAMS) have been observed (Schwartz et al., 1992). The quasi-parallel bow shock is replaced repeatedly by newly forming shocks, which leads to strong fluctuations also in the magnetosheath downstream (Blanco-Cano et al., 2006a, 2006bBurgess, 1989;Scholer & Burgess, 1992;Schwartz et al., 1992).
During the last two decades, a number of authors have reported observations of plasma entities that stand out from the surrounding magnetosheath by having either an enhanced density, speed, or both. A few different terms have been used to denote these structures, for example, "magnetosheath dynamic pressure enhancements" (Archer & Horbury, 2013), "density enhancements" (Gutynska et al., 2015), "transient flux enhancements" (Nĕmeček et al., 1998), "antisunward high-speed jets" (HSJ) , "supermagnetosonic subsolar magnetosheath jets" (Hietala et al., 2012), "high kinetic energy density plasma jets" (Savin et al., 2008), "large-scale jet" (Dmitriev & Suvorova, 2015), and "super fast plasma streams" (Savin et al., 2012). However, Gunell et al. (2014) used the term "plasmoid" to describe velocity structures with the typical scales on the order of 1 Earth radius (R E ). The same term was used by Karlsson et al. (2012) to investigate enhancements in the magnetosheath density. In spite of the disparate terminology, one could, in a general sense, treat these terms as synonyms. However, the detailed properties of the entities observed do depend on precise definitions and selection criteria. We shall use the term "jet" for all the plasma entities in our data set and "fast plasmoid" for the subset whose elements also show a speed increase of 10% or more. The criteria used in this work and in some of the previously published studies are summarized in Table 1.
Statistical studies have shown that jet occurrence is almost exclusively controlled by the angle between the IMF and the Earth-Sun line (cone angle), while other solar wind parameters or their variability only play a minor role. The jets are predominantly observed when this cone angle is small, that is to say, downstream of the quasi-parallel shock (Archer & Horbury, 2013;Plaschke et al., 2013). It was also found that 97% of the observed jets can be formed locally at the bow shock-as opposed to upstream in the solar wind-by ripples that appear on the shock when it is quasi-parallel .
In contrast to the quasi-parallel bow shock, the quasi-perpendicular shock is well-defined, with a relatively thin ion ramp (thickness of several thermal gyroradii), and the magnetosheath is less turbulent downstream of the quasi-perpendicular than the quasi-parallel bow shock. The reflected ions gain enough energy to pass through the shock front from only one full gyromotion, and there is not enough time for waves to grow. Nevertheless, locally generated waves have been observed in the downstream region (Goncharov et al., 2014;Hoilijoki et al., 2016;Lembège et al., 2004;Mazelle et al., 2003;Nĕmeček et al., 2013;Ofman et al., 2009;Ofman & Gedalin, 2013;Soucek et al., 2015;Yang et al., 2012;Yang, Lu, Lembège et al., 2009). Recently, shock ripples have been observed at quasi-perpendicular bow shock (Johlander et al., 2016). Johlander et al. (2018) concluded that the shock ripples propagated along the shock surface as quasi-periodic fluctuations with Alfvénic wave-like structure. Gutynska et al. (2015) performed an analysis of the wave properties for understanding the nature of the magnetosheath plasma structures and concluded that events with significant density and pressure enhancements are fast magnetosonic waves. Global electromagnetic hybrid simulations (Omidi et al., 2014) suggest that jets that have been formed at the quasi-parallel bow shock for small IMF cone angles may extend into the quasi-perpendicular magnetosheath at the flanks.
On the other hand, at the flanks of the magnetosheath, the magnetosheath plasma stream is itself supermagnetosonic, and bow shock ripples would not necessarily create discernible jets. Archer and Horbury (2013) reported that jets become less common toward the flanks. However, Karlsson et al. (2015) observed density enhancements throughout the dayside flanks. Despite the increasing number of observational studies, the jet formation mechanism remains an open question. Karlsson et al. (2015) suggested that SLAMS (Schwartz, 1991;Schwartz et al., 1992) could transform into jets when traveling through the bow shock. It was found in a global hybrid-Vlasov simulation that a jet can be created through the interaction between the bow shock and a SLAMS-like structure passing through it (Palmroth et al., 2018). At least a 50% increase of n e and less than 10% increase of v above a 500-s running average Karlsson et al. (2012) Fast plasmoid At least a 50% increase of n e and at least 10% increase of v above a 500-s running average Karlsson et al. (2012)

Diamagnetic plasmoid
At least a 50% increase of n e (30% in solar wind) and ΔB < 0 (ΔB: maximum deviation from a 5-10-min average before or after plasmoid encounter) Karlsson et al. (2015) Paramagnetic plasmoid (subset of magnetosheath jets) In this paper, we investigate jets in the quasi-perpendicular magnetosheath. We will compare their evolution and relation to upstream parameters, with previous statistical studies of jets, downstream of the quasi-parallel bow shock. We use data from the four Magnetospheric Multiscale (MMS) spacecraft , orbiting in the magnetosheath and bow shock region, and NASA OMNI high-resolution solar wind data (King & Papitashvili, 2005), gathered during the years 2015-2017. We compare several models of the jet formation mechanism and provide quantitative predictions of jet propagation toward the magnetopause.

Data Set and Methodology
From October 2015, the orbits of MMS spacecraft cover a broad region of the magnetosheath. Our visual inspection of the ion energy spectra, plasma, and magnetic field signatures in MMS Quicklook plots (https://lasp.colorado.edu/mms/sdc/public/plots/quicklook) excluded the magnetosphere and solar wind parts. We used fast survey mode resolution of the magnetic field (16 Hz) and plasma (4.5 s/sample) measurements (Pollock et al., 2016;Russell et al., 2014). The electric current density, J, was estimated by the curlometer method from the four MMS magnetic field data (Dunlop et al., 2002;Paschmann & Schwartz, 2000).
Different physical parameters have been used to identify jets previously. The choice of parameters often affect the nomenclature and the numbers of selected events. The best choice of an identification criterion to use in future studies is still under debate, and it will depend on the particular science questions and the availability of measurements. In this work, we compare several criteria, which have been adopted from previous observational studies. Each selected interval of the magnetosheath region was automatically analyzed to look for the enhancements in the antisunward (negative X component) velocity (Gunell et al., 2014;Hietala et al., 2009Hietala et al., , 2012. All events were visually inspected and compared with the Plaschke et al. (2013) (P13) identification criterion, defined by the ratio of the magnetosheath dynamic pressure in the X direction (P dx = ρ·V x 2 ) to the upstream solar wind dynamic pressure (P dsw = ρ·V sw 2 ), where ρ is the mass density.
According to their selected criterion, the jet is a region where P dx > 0.25·P dsw with a maximum value higher than 0.5·P dsw . Furthermore, this criterion would only be applicable in the subsolar region of the magnetosheath since the magnetosheath velocity in the flanks increases with respect to the solar wind velocity, and this inequality would be satisfied almost all the time. A summary of these criteria can be viewed in Table 1.
A comparative analysis of different criteria by  predicted that the Archer and Horbury (2013) (AH13), P13, and Karlsson et al. (2012Karlsson et al. ( , 2015 (K12) criteria identified the most events and that of these, the lower threshold of AH13 provide the highest number of events. However, the actual overlap in the three definitions, which were applied to the Archer and Horbury (2013) survey with no further assumptions, showed that just under 3% of all identified data points are common to them all. About 60% of all observations identified by K12 criterion also satisfy one of the other two definitions. On the other hand, about half of the samples identified by P13 are not satisfied by the other two definitions. The study of the overlap between P13 and AH13 definitions on the Plaschke et al. (2013) survey showed agreement only in 17% of the events (Vuorinen et al., 2019). The authors suggested that the AH13 criterion includes not only jets but the numbers of the other structures close to the magnetopause.
The criterion used in this work is a combination of the AH13 and P13 criteria. According to this, a jet event is a time interval where P dx is greater than 0.4·P dsw for all measurement points and the maximum P dx in this interval is greater than P dsw . As a result, 1,400 jets are selected based on these criteria. According to AH13, jets are only considered when the maximum dynamic pressure (P d ) of the jet is greater than twice the 20-min temporal average of the surrounding magnetosheath dynamic pressure. Our selected events show excellent agreement with this criterion.
Localized structures with an electron density of at least 50% higher than the surrounding plasma were termed as plasmoids by Karlsson et al. (2012). According to Gunell et al. (2012Gunell et al. ( , 2014, plasmoids are associated with an increase in flow velocity. Plasmoids with velocity changes of more than 10% were termed fast plasmoids by Karlsson et al. (2012Karlsson et al. ( , 2015. Only half of the selected events in our study fully meet this criterion. In the present paper, we refer them as "fast plasmoids." The upstream solar wind parameters (i.e., magnetic field, velocity, density, and temperature) were obtained from the 1-min OMNI solar wind data, propagated to the bow shock nose and consisted of measurements from different spacecraft around the L1 point. For instance, to compensate the uncertainty of the solar wind propagation from the spacecraft at the L1 point to the bow shock as well as to the jet locations, all upstream solar wind parameters were calculated by averaging the data 4 min before and after the observations of jets. To obtain the minimal distance between the bow shock and the spacecraft as well as the local bow shock normal, the calculated solar wind parameters were compared with the Farris and Russell (1994) bow shock model.

Observations
Early studies of jets in the magnetosheath have indicated that the jet formation is closely related to processes in the foreshock region of quasi-parallel bow shock. Our data set shows that jet occurrence downstream of the quasi-perpendicular shock is not uncommon. However, one should note that we have used a somewhat different selection criterion, and the jets in our data set show higher antisunward velocity enhancements. Figure 1a presents the distributions of the θ Bn and cone angles at time of jet observations, represented by red and blue colored lines, respectively. Figure 1a shows that such jets are observed in a wide range of θ Bn and cone angles but are primarily observed during typical Parker's spiral IMF (i.e., cone angles~45°). However, the normalized distribution of cone angles (green line) shows that jets occur much more often downstream of the quasi-parallel bow shock.
The θ Bn angle calculated at the closer bow shock point provides actual information about the magnetosheath, where jet propagated. According to previous studies, the velocity of magnetosheath jets is generally oriented more along the Sun-Earth line (Archer & Horbury, 2013;Gunell et al., 2012;Hietala et al., 2012;Plaschke et al., 2013). Figure 1b presents the relationship between the distance along the Sun-Earth line and the minimal distance to the bow shock in geocentric solar ecliptic (GSE) coordinate system. Due to the orbits of the MMS spacecraft during 2015-2017, the magnetosheath intervals are  (Farris & Russell, 1994) and magnetopause (Shue et al., 1998). Color coding corresponds to distance to the bow shock.
mainly observed in the subsolar magnetosheath, and the differences between these distances are small. However, jets are connected not only with the increasing of the X component of the velocity flow. The values of the Y and/or Z components can be up to 70% of X, and they can affect the jet propagation direction. For this reason, in our analysis, we use the minimal distance to follow the jet propagation.
On the other hand, the θ Bn angle highly affects the propagation of the jets. Figure 1c presents a map of the jet in the GSE coordinate system. In this figure, the locations of the jets are separated based on whether they are observed in the regions of quasi-parallel or quasi-perpendicular magnetosheath. Jets that are observed downstream of quasi-parallel (quasi-perpendicular) shock are plotted in the positive (negative) R GSE region. The colored dots correspond to different ranges of the minimal distance from the jets to the model bow shock locations. It is seen in Figures 1b and 1c that jets can be observed in a wide range of distance to the bow shock, but they are predominantly found closer to the bow shock (distance from the bow shock <2 R E ) than to the magnetopause. A similar result was also noted by Plaschke et al. (2013). However, it is noteworthy to mention that jets downstream of the quasi-perpendicular shock are observed only up to 2.5 R E . Palmroth et al. (2018) used a global hybrid simulation and found that the jet dynamic pressure is greatest nearest to the shock and decreases as the jets propagate toward the magnetopause. The dynamic pressure decreases by 70% from the bow shock to the vicinity of the magnetopause. This effect indicates that the origin of the jet may be related to the dynamic pressure upstream the bow shock.
Our statistical analysis is based on MMS observations in the magnetosheath. One thousand four hundred jets were selected in this work, and as shown in Figure 1b, about 35% of the events were observed downstream of the quasi-perpendicular bow shock. One example of such events is shown in Figure 2.  Figure 2a is always greater than 50°, suggesting that the spacecraft is located downstream of the quasi-perpendicular bow shock. In addition, the rather narrow ion energy range shown in Figure 2c is a typical characteristic of the quasi-perpendicular magnetosheath. Furthermore, according to the Farris and Russell (1994)  In a case study, Eriksson et al. (2016) observed a high current density inside a jet and concluded that a current sheet had formed at the boundary between colder, more solar wind-like plasma and warmer, magnetosheath-like plasma. The current sheet, as well as the jet, propagated from the bow shock toward the magnetopause. Detailed analysis of jet by Plaschke et al. (2017) showed that jet can have a significant substructure, including the current sheet and waves. Similar to previously reported jets which predominantly have been observed in the quasi-parallel magnetosheath (e.g., Eriksson et al., 2016;Plaschke et al., 2013Plaschke et al., , 2017, the jets we identify in this work, downstream of quasi-perpendicular shock, are also 10.1029/2019JA027667

Journal of Geophysical Research: Space Physics
GONCHAROV ET AL.  (Figure 2g). In Figure 2j, it is seen that the ion energy flux is high inside the jets, peaking at an energy of~0.8 keV, which is not far below the typical solar wind proton energy of~1 keV. However, these signatures can also be observed for "small jets" that do not match any of the criteria discussed in this work, for example, in the approximate period from 10:17:36 to 10:17:46 UT as shown in Figure 2j. Hence, jets should not be identified based solely on one single parameter. A comparison of the energy flux with signatures in other measured quantities is more likely to provide a good identification criterion.
Based on the magnetic field strength compared to the surrounding background field, Karlsson et al. (2015) divided the plasmoids they observed into two distinct groups. Plasmoids that were associated with clear magnetic field increases were called "paramagnetic," and plasmoids with clear decreases were called "diamagnetic." The majority of the jets and fast plasmoids identified in our data set have paramagnetic signatures. However, it is of interest to note that the magnetic field signatures within the jets and fast plasmoids themselves are not always same. For instance, it is shown in Figure 2e that the magnetic field of the jets (boxes with blue dashed lines) are accompanied with larger fluctuations and more rapid rotations compared to the fast plasmoids (boxes with yellow dashed lines). However, a more detailed analysis with a larger number of examples is needed and will be done in the near future.
On the other hand, the changes in the ion and electron densities of the jets and plasmoids show excellent anticorrelation with the changes in ion temperature (comparing Figures 2h and 2k to Figure 2i). Such a signature is observed in all 1,400 cases (results not shown) and is consistent with the study by Gutynska et al. (2015), who analyzed jets based on density enhancements using a K12 criterion. An anticorrelation between density and temperature is a typical signature of magnetosonic waves. In other words, these results suggest that jets and fast plasmoids are magnetosonic in nature.

Upstream and Downstream Conditions
Apart from the clear dependence on cone angles, Plaschke et al. (2013) pointed out that jets in the quasi-parallel magnetosheath have rather weak to insignificant dependence on solar wind velocities and magnetic field strengths. Figure 3 shows histograms of the upstream solar wind parameters, (a-d)

10.1029/2019JA027667
Journal of Geophysical Research: Space Physics magnetic field components and magnitude, (e) plasma velocity magnitude, (f) dynamic pressure, (g) Alfvén Mach number (M A = |V|/V a , where V a = B j j= ffiffiffiffiffiffiffi μ 0 ρ p is the Alfvén speed and μ 0 the vacuum permittivity), and (h) plasma beta (β = 2μ 0 k B NT/|B| 2 , where k B is Boltzmann's constant, N is the solar wind density, and T is the solar wind temperature), at the time when the jets were observed in quasi-parallel (green lines) and quasi-perpendicular (red lines) magnetosheath. In all panels, the blue lines show the respective parameters in the entire time periods regardless of whether jets were identified. Figures 3a to 3c show that there is no (clear) dependence between the jet occurrence and the IMF orientations. In Figures 3d and 3e, the difference between the blue and green distributions shows that the occurrence of jets in the quasi-parallel magnetosheath has clearer dependence on the magnitude of both the magnetic field and plasma velocity. In addition, our results show that the occurrence of jets in the quasi-perpendicular magnetosheath is linked to higher (about 20%) solar wind magnetic field and plasma velocity magnitudes than jets in the parallel magnetosheath. Similar dependence can also be observed in the dynamic pressure and Alfvén Mach number shown in Figures 3f and 3g. Furthermore, in contrast to Plaschke et al. (2013), the occurrence of jets in our work shows strong dependence on the solar wind plasma beta. In particular, higher plasma beta is observed when jets are present in the magnetosheath. The plasma beta is 30% higher when the jets are observed in the quasi-perpendicular magnetosheath compared to quasi-parallel magnetosheath (red and green lines in Figure 3h). Figure 4 shows the change in direction of the magnetic field and the plasma bulk velocity between the solar wind, the magnetosheath, and the jets. Figure 4a shows a histogram of the angle between the velocity vector in the solar wind and in the magnetosheath, and Figure 4b shows how this angle varies with the distance to the bow shock. Histograms of the angle between the magnetic field in the solar wind and in the magnetosheath and how it changes with the distance to the bow shock are shown in Figures 4c and 4d, respectively. In all panels, the quasi-parallel magnetosheath is shown in green and the quasi-perpendicular in red. The 95% confidence intervals of the medians are shown by colored dotted lines in the first and fourth columns and by colored area in the second and third columns. The changes of the IMF orientation in the quasi-perpendicular magnetosheath are smaller (red line in Figure 4d) than in the quasi-parallel magnetosheath. Dimmock and Nykyri (2013) showed that the quasi-perpendicular magnetosheath is a region favorable to magnetic field and velocity asymmetry (i.e., velocity and magnetic field strength are larger on the dusk flank) during typical Parker spiral IMF conditions. However, deflections of the plasma flow in the quasi-parallel and quasi-perpendicular magnetosheaths are almost the same (Figure 4a). The triangles and circles in Figures 4b and 4c show that the deflection of the IMF and plasma flow from the solar wind orientation increases with the propagation depth in the sheath. Due to higher fluctuations in the quasi-parallel magnetosheath, calculation errors of the magnetic field rotation are rather large (green triangles).
Previous studies concluded that the velocity of magnetosheath jets is generally oriented more along the Sun-Earth line (Archer & Horbury, 2013;Gunell et al., 2012;Hietala et al., 2012;Plaschke et al., 2013). We find that the median deflection of the solar wind flow from the Sun-Earth line is 2.68°(not shown). Figures 4e to 4h show the changes of the jet plasma flows, as well as the magnetic fields, from the solar wind orientation. The colored circles in Figure 4f show that the deflection angle increases with bow shock distance, similar to the magnetosheath plasma in Figure 4b. However, the jets are not deflected as much as the rest of the magnetosheath. For instance, the deflection angle between the jet and the magnetosheath plasma is about 15°, as shown in Figure 4i. Gunell et al. (2012) concluded that the plasmoids move predominantly in a tangential direction with respect to the magnetopause normal. On the other hand, Hietala and Plaschke (2013) reported that the jets have a tendency to continue "straight" along the Sun-Earth line as compared to the background magnetosheath flow. The deflection from the background magnetosheath flow is in the range of 20°to 45°. Plaschke et al. (2013) reported a number of 28.6°for the median deflection. In our statistics, the median deflection for both types of jets are quite similar and in a good agreement with the early reported range (Figures 4i to 4l). Thus, we conclude that the direction of propagation is in between the extremes of being tangential to the magnetopause and continuing straight in the solar wind direction. Figures 4j to 4k show that differences of the magnetic field (triangles) and flow (circles) orientations inside and outside of the jet slightly decreases with propagation depth in the quasi-parallel magnetosheath. Nevertheless, in the quasi-perpendicular magnetosheath, the deflection angle is almost constant with the propagation depth, and this confirms the hypothesis that the jets move in their original direction with respect to the magnetosheath.

Journal of Geophysical Research: Space Physics
Propagating through the magnetosheath, jets do not only affect the magnetopause and magnetosphere. Simulation results by Karimabadi et al. (2014) showed that jets pushed slower ambient magnetosheath plasma out of their way. As a result, the plasma moves around the jets, and it is slowed down or could even be pushed in the sunward direction. Consequently, jets may create anomalous flows and be a source of additional turbulence. Figure 4m shows a histogram of the angle between the magnetosheath flow direction ahead and behind the jets. In this work, the majority of the jets are associated with small changes in the magnetosheath flow. However, from Figure 4p, it can be seen that jets are accompanied by strong changes in the presents changes of the solar wind parameters in the magnetosheath (SW&MSH) and inside jet (SW&Jet). Differences of the magnetosheath parameters inside (MSH&Jet) and around (MSH pre&post ) jets are presented in panels (i)-(l) and (m)-(p), respectively. In all panels, the green and red colors represent the quasi-parallel and the quasi-perpendicular magnetosheath, respectively. Filled circles correspond to velocity deflections and filled triangles to magnetic field deflections. The 95% confidence intervals of the medians are shown by colored dotted lines in the first and fourth columns and by colored area in the second and third columns.

10.1029/2019JA027667
Journal of Geophysical Research: Space Physics magnetosheath magnetic field orientation. On the other hand, the influence of the jets on the surrounding magnetosheath decreases with the propagation distances (Figures 4n and 4o).
The magnetosheath jet velocity is considerably greater than the local Alfvén velocity. Some jets are also supermagnetosonic and may even be associated with a local shock at the front of the jet (Hietala et al., 2009(Hietala et al., , 2012Plaschke et al., 2017;. In our analysis, about 95% of jets are superalfvénic (not shown) with median Alfvénic Mach number of 1.8 in the quasi-parallel and 1.6 in the quasi-perpendicular magnetosheath. Nevertheless, only 12% of them are supermagnetosonic (i.e., jet speed is higher than magnetosonic speed, V ms = is the sound speed, γ is the adiabatic index). Plaschke et al. (2013) noted that in the subsolar region, only about 14% of jets are supermagnetosonic. Archer and Horbury (2013) suggested that the majority of the supermagnetosonic jets must be observed in the flanks of the magnetosheath. On the other hand, they pointed out that density-driven jets are more likely to occur at the flanks. Covering the whole dayside region, Archer and Horbury (2013) reported that there is no clear change in observation probability with distance from the bow shock. The analysis of plasmoids in the magnetosheath by Karlsson et al. (2015) showed that plasmoids with changes in velocity are found for X GSE > 2 R E , while plasmoids without velocity changes (labeled embedded plasmoids by Karlsson et al., 2015) are found further downstream for X GSE > −5 R E . Figures 1b, 1c, 4f, and 4j show that velocity-driven jets are more common close to the bow shock and deflections from the magnetosheath flow decrease with increasing propagation distances. Such an effect can be the result of jets slowing down while propagating through the magnetosheath. Figure 5 shows the ratio of jet speed to magnetosheath flow speed as a function of bow shock distances for jets (a, b) and fast plasmoids (c, d) in the quasi-parallel and quasi-perpendicular magnetosheaths.
In the quasi-parallel magnetosheath as well as in the quasi-perpendicular, the median values in each 0.5 R E distance bins show a signature of the jets and fast plasmoids slowing down. Similar signatures were observed by Dmitriev and Suvorova (2015), who followed jets with the five THEMIS spacecraft and reported a decrease of the jet velocity as it moved toward the magnetopause. Differences of the deceleration trend in the quasi-perpendicular and quasi-parallel magnetosheath can be connected with the formation mechanisms and the asymmetry of the magnetosheath magnetic field and velocity, associated with a quasi-perpendicular bow shock (Dimmock & Nykyri, 2013).

Discussion
We have presented a statistical study of magnetosheath jets and fast plasmoids in the quasi-parallel and quasi-perpendicular magnetosheaths. We compared the main observational criteria used in the past (Archer & Horbury, 2013;Karlsson et al., 2012Karlsson et al., , 2015Plaschke et al., 2013), and we could verify the properties of jets and fast plasmoids. Our results mainly confirmed previous statistical studies and show good agreement between jets and fast plasmoids. We showed that jets occur not only downstream of the quasi-parallel shock and that jets in the quasi-perpendicular magnetosheath are not rare. We found that the occurrence of jets is not exclusive to low IMF cone angles and they can be detected even during a perpendicular IMF orientation. The existence of jets in the quasi-perpendicular magnetosheath is mainly connected with oblique IMF orientations. A strong solar wind with higher than average velocity and magnetic field (i.e., higher Alfvén Mach number and plasma beta) creates conditions favorable for the generation of jets.
As they propagate through the magnetosheath, the jets accelerate slower plasma ahead of them in their propagation path and push it out of their way , and they are slowed down in the process. Dmitriev and Suvorova (2015) discussed a large-scale jet whose velocity decreased about 20% over a 1.4 R E propagation distance in the magnetosheath. The trends in Figure 5 show a quite fast deceleration rate in both the perpendicular and parallel magnetosheath. According to the best fit equation in our results, at distances of about 10-11 R E from the bow shock, jets will have acquired the same velocity as the surrounding magnetosheath. The jets and fast plasmoids in the quasi-perpendicular magnetosheath have shown much 10.1029/2019JA027667

Journal of Geophysical Research: Space Physics
faster deceleration and at distances of about 4-5 R E will be embedded in the magnetosheath flow. Such fast deceleration of jets and plasmoids in the magnetosheath can explain the results of Karlsson et al. (2015), who pointed out that plasmoids were observed up to X GSE = −5 R E , but fast plasmoids only until X GSE = 2 R E . Figures 4j and 5a indicate that the propagation of jet through the magnetosheath since their formation is not constant. This may provide hints on their propagation evolution. It is clear seen that tangential flows are significant and not negligible. Hence, it remains a question whether the jets in their later stages (deeper in the magnetosheath when they may have lost their initial propagation momentum) would still propagate mainly in their nominal directions or flow along the now-ambient sheath flow. Further work should be done to validate this hypothesis.
What mechanism leads to the formation of jets is still under debate. Many different models have been suggested. The majority of these conclude that foreshock processes are responsible for generating most of the jets (Hietala et al., 2012;Plaschke et al., 2013). Among the more notable of the suggested mechanisms, we find bow shock ripples and SLAMS, which are both inherent to the quasi-parallel shock, although ripples have also been observed on the quasi-perpendicular shock (Johlander et al., 2016(Johlander et al., , 2018. On the other hand, Archer et al. (2012) suggested that jets could form when the shock locally changes from quasi-parallel to quasi-perpendicular or vice versa due to IMF discontinuities passing the bow shock. However, jets are mostly observed during steady IMF, and only 15% of the jets in this study were connected with a clear signature of a solar wind discontinuity. For this reason, alternative formation mechanisms are also needed to explain the observations.
Early estimates of the size of the magnetosheath jets gave values in the order of 1 R E (Archer et al., 2012;Gutynska et al., 2015;Hietala et al., 2009Hietala et al., , 2012Karlsson et al., 2012Karlsson et al., , 2015. However, reports of the detailed morphology of the jets have been somewhat inconsistent. Archer et al. (2012) report a longer scale size parallel to the jet flow than perpendicular to it. Plaschke et al. (2016) interpreted their results as the jets having a

10.1029/2019JA027667
Journal of Geophysical Research: Space Physics pancake-like geometry. Similar to previous studies, we assumed a cylindrical jet geometry. The dimension along the flow direction (D||) was estimated for every jet and fast plasmoid observation by integrating the ion velocity over the jet duration (Δt). The duration times were determined as the time of jet observation, where the P dx is greater than 0.4·P dsw . An upper limit of the flow parallel size was proposed by Gunell et al. (2014) as the product of the duration and maximum speed, and this provide excellent agreement with D|| in the all distance ranges (not shown).
To determine thickness (Dp) of jets, we use the Rankine-Hugoniot relations (Koval & Szabo, 2008) to obtain the local normal and speed along this estimated normal. In particular, we require the angle between the obtained normal and the jet flow direction to be in the 80°to 100°range. Applying these criteria, we obtain a set of 980 cases of jet and fast plasmoid observations in the quasi-parallel and quasi-perpendicular magnetosheaths. Assuming that the jet propagated along the normal with the estimated speed, the jet thickness was calculated. The thickness is an overestimate as we have not taken into account propagation parallel to the flow (V jet ). Figure 6a shows a sketch of our assumption. The spacecraft is located on top, and the estimated trajectory of its motion is shown by dashed colored lines. Figures 6b and 6c show the shapes of the jets (blue) and fast plasmoids (green) in the quasi-parallel (b) and quasi-perpendicular (c) magnetosheaths. The light blue and light green lines, in both figures, represent the median values of the jet and fast plasmoid D|| sizes at different distances from the bow shock. The dotted rectangles in panels (b) and (c) represent the median size of the jets (blue) and fast plasmoids (green) in both types of the magnetosheath.  dotted rectangles, respectively. All plots indicate continuous increasing trends of the parallel, D|| (upper panels), and perpendicular, Dp (lower panels), dimensions of jets in the quasi-parallel (left) and quasi-perpendicular (right) magnetosheath. The median values show that the extent of both jets and fast plasmoids parallel to the direction of propagation is almost twice as high as the perpendicular extent irrespective of magnetosheath type. A similar conclusion was reported by Archer et al. (2012). However, the slopes of the curves described by the red circles indicate that the jets expand faster in the perpendicular than the parallel direction as they travel through the magnetosheath. Similar results were observed for fast plasmoids (not shown) and confirmed the conclusion by Karlsson et al. (2015) that the plasmoids are a subset of magnetosheath jets.

Summary and Conclusion
Based on our analysis of 1,400 events, with higher enhancements in the X component of the velocity, in a broad range of the magnetosheath and upstream parameters, we can summarize our conclusions in the following list.
1. The basic properties of jets and fast plasmoids are very similar. 2. The probability of jet generation increases when the solar wind is stronger, that is, when it has a higher velocity, magnetic field, Alfvén Mach number, and plasma beta. 3. Jet observations in the quasi-perpendicular magnetosheath are relatively common. 4. Jet observations are more frequent close to the bow shock, and their direction is toward the magnetopause. 5. A low θ Bn angle enables propagation deeper into the magnetosheath. 6. The propagation speed of a jet decreases as it moves toward the magnetopause, and during its propagation, the velocity of the jet tends toward the velocity of the surrounding magnetosheath. 7. The typical size of these structures is several thousands of kilometers, and it increases with the distance to bow shock. 8. The parallel size of the jets and plasmoids is almost two times higher than perpendicular size in both the parallel and perpendicular magnetosheaths.
Our comparative analysis showed no significant differences between the plasma properties of the jets and fast plasmoids. However, the different magnetic fields inside the structures suggest that the formation mechanisms are different. Palmroth et al. (2018) found that the criteria applied by Archer and Horbury (2013) and Karlsson et al. (2012) provide a better opportunity than the criterion by Plaschke et al. (2013) to detect jets "shaped more like blobs." Further comparative analysis of the detailed structure of these plasmoids and jets is necessary and will be conducted in the near future.