NOAA Polar-orbiting Operational Environmental Satellite (POES)

An explanation of the POES mission; its data gathering and interpretation process as a tool for amateur radio operators in analyzing HF radio propagation

Written for the BCDX Club by Dave S. Evans, the Space Environment Center scientist responsible for the NOAA/POES Space Environment Monitor data process.

In late 2002 the BCDX Club made inquires with NOAA (National Oceanic Atmospheric Administration) in Boulder, CO regarding the Auroral Activity imagery Extrapolated from NOAA POES at http://sec.noaa.gov/pmap/ . In the course of our inquiry, we met David S. Evans -- the scientist directly responsible for the Auroral data gathering process at NOAA. Through our telephone conversations and ensuing email exchange, we began a discussion about the Auroral impact on HF radio here in the Pacific Northwest area of North America -- specifically British Columbia. 

The net result of our exchange is the following essay written by David S. Evans for the BCDX Club. We share the material here in hopes it will help radio operators better understand how to interpret the Auroral Activity data  Extrapolated from NOAA POES satellite. The BCDX Club wishes to thank David Evans for the time and energy he has devoted to this project. To position the expertise of our author, we offer you the following personal snapshot:

Introduction to the Auroral Activity Extrapolated from NOAA POES

Instruments on board the NOAA Polar-orbiting Operational Environmental Satellite (POES) continually monitor the power flux carried by the protons and electrons that produce aurora in the atmosphere. SEC has developed a technique that uses the power flux observations obtained during a single pass of the satellite over a polar region (which takes about 25 minutes) to estimate the total power deposited in an entire polar region by these auroral particles. The power input estimate is converted to an auroral activity index that ranges from 1 to 10.

Over 100,000 satellite passes comprise the NOAA POES historical database that was used to construct statistical patterns of auroral power flux for each of the 10 levels of auroral activity as defined by total power dissipation as illustrated in the following table.

Note from the author:  This essay probably rambles but I wanted to cover all I thought important. Included are the nature of the Space Environment Monitor (SEM instruments on the satellites), the nature of the observations we make and how we manipulate the data, the nature of the satellite orbits and the data recovery process (which determine many aspects of the way the data are presented), and finally how to extract the maximum of information from what we present.  A lot of these points will probably be mixed together and I apologize if things are not as coherent as they could be.  DE

Starting with some definitions

We tend to think in terms of magnetic latitude and magnetic local time.  To make life confusing right from the start, magnetic latitude comes in at least two flavors.  The first is simple dipole magnetic latitude, which is determined from complement of the great circle arc length between the geographic location in question and the magnetic dipole north pole.  That pole is currently located near 79.6 N geographic latitude, 288.3 E geographic longitude around the NW coast of Greenland.  (Incidentally, there are several ways to define the magnetic pole – the “dipole” pole in Greenland, the “dip” pole that is further west in the Canadian archipelago, and the “invariant” pole which is somewhere is between.)  The second magnetic latitude we use is “corrected magnetic latitude” (CML) which is more complex as it takes into account the higher order, non-dipole terms in the Earth’s magnetic field.  A CML of 90º is up at about 81.5N, 280E geographic.  While we use both forms of magnetic latitude, the CML is better for dealing with auroral phenomena. 

Magnetic local time (MLT) is much like local solar time, except the Earth’s magnetic axis replaces the Earth’s axis of rotation is setting up the geometry.  At geomagnetic latitudes of more than 30º or so from the magnetic pole, there is not much difference between MLT and local solar time – perhaps 45 minutes.  However, near the magnetic poles there can be very large differences.  For example, if you were located between the magnetic pole and the geographic pole, it could be midnight, local solar time, but noon MLT.  Magnetic local time orders auroral data much better than solar time and so we use it. 

We also occasionally refer to L-value.  L-value is a rather arcane concept first developed to “label” magnetic field lines in order to better understand observations of the Van Allan radiation belts.  Fortunately, L-value has a fairly simple geometric interpretation.  It is about the distance, in units of Earth radii, from the center of the Earth that the magnetic field line emanating from a given CML location will cross the magnetic equator.   In space, boundaries between particle populations usually follow an L-value.  For many of our displays we plot L-value contours because it helps the viewer relate an observation at one location to what be happening at another.  To get a little mathematical,

L-value =1.0/(cos(CML))². 

That is, if the CML is 45º, L-value=2.0 (the field line will loop out to 2 Earth radii from the center of the Earth before returning to the opposite hemisphere), if 60º, L-value=4.0 and so on. 

We tend to use electron Volts (eV) as a unit of particle energy.  One electron Volt is the kinetic energy acquired by a charged particle that has been accelerated through one-Volt electric potential.  The beam of electrons that create the color picture in a TV set have about 25,000 eV of energy, somewhat higher energy than the electrons responsible for aurora. 

We also refer to the ‘foot of the field line’ (fofl).  What we do is trace along the magnetic field line from the satellite (where the particles are observed) down to 120 km altitude where the particles actually enter the atmosphere.  It is the latter location that is more relevant to knowing where the aurora is actually occurring and where the ionosphere is disturbed. 

The instruments on POES

The Space Environment Monitor (SEM) package on the POES satellites contains a total of 16 separate particle detectors, but they break down into 2 areas.  

The first is the Total Energy Detector (TED) that concentrates on observing the particles that produce aurora.  The eight TED sensors have an electrically controlled ‘particle energy filter’ followed by a detector that counts each particle that makes it through the filter.  The range of particle energies (and the separation between electrons and protons) passing through the filter is selected by changing a voltage.  On the POES satellites we step this ‘energy filter’ through 16 steps every two seconds.  The particle energies at the center of each of the 16 energy bands are listed below.  

Similar detector systems are flown on scientific satellites and return very detailed information about both the energy and angular distributions of those particles that are precipitating into the atmosphere to create aurora.  However, we were given very limited telemetry on the POES satellites and had to make decisions about what observations we really needed.  We focused on measuring the integrated flux energy flux carried by particles that would end up in the atmosphere rather than trying to get all the details.  We felt that this energy flow would be a good guide to the level of activity as well as giving a good idea of the level of disturbance created in the atmosphere below when these particles were stopped.  As a secondary measurement we also identified which of the 16 energy bands contained the maximum in the energy flux distribution, as this would give an idea of how deep in the atmosphere the incoming particles would penetrate.  

Telemetry limitations make it impossible to send back the observations from all detectors at each of the 16 energies and do the arithmetic in the ground processing.  There is a microprocessor on the satellite that manipulates the data to obtain a measure of the total energy flux summed up over all the particle energies observed and the result of that calculation is transmitted to the ground.  We make a pair of measurements, for electrons and protons, at each of two angles with respect to the magnetic field direction at the satellite.  Every 2-seconds we get a set of 8 primary observations (2 energy flux measurements and 2 energy band measurements for electrons and for protons) of the auroral particle flow towards the atmosphere.  There are additional data obtained at a lower duty cycle, but those data are not of interest here. 

The following is a table of particle energy at the center of each energy band and the altitude in the earth’s atmosphere that an electron of that energy will produce maximum ionization rates and the altitude they will stop.  Protons of the same energy will be stopped somewhat higher in the atmosphere than electrons of the same energy.  I have included two higher energies as well because additional detectors on the satellite measure electrons at those energies.

The 16 energy channels listed above map to the number of dots (1 to 16) on our web display.  As one can see, if the incoming electrons and protons have energies less than about 1000 eV or so, the ionization they produce in stopping is mainly above 150 km.  The increased electron densities at those altitudes might affect HF signal propagation paths but will not produce any significant signal absorption.  As the particle energy gets higher and the ionization produced at lower altitudes, absorption in the E-layer and even the D-layer (if the 100,000 eV electron fluxes are high) might be significant. 

The other parameter that is returned from these detectors is the net energy flux integrated between 50 eV and 20,000 eV.  This number is returned separately for the proton and electron measurements and for each of the two viewing angles.  In ground processing we sum the electron and proton directional energy fluxes and perform an integration over angle to get the ‘total’ energy flux into the atmosphere carried by all particles of energies between 50 and 20,000 eV – hence the name Total Energy Detector. 

The range of energy fluxes the instrument measures goes from about .001 mW m^-2 up to about 500 mW m^-2.  For point of reference: 

• 0.25 mW m^-2 will produce an aurora just visible to the dark-adapted eye under good seeing conditions (no moon, clear sky.)

• 2.5 mW m^-2 will produce an aurora easily visible given good seeing conditions.

• 25 mW m^-2 will produce an aurora visible to the non-adapted eye under poor seeing conditions, for example immediately upon leaving the pub.

• 250 mW m^-2 will produce an aurora that you could almost read a newspaper by. 

We have measured up to 400 or 500 mW m and on very rare occasions I suspect the energy flux to the atmosphere could reach 1000 mW m^-2. 

The 2-second cadence for making the total energy flux observation converts to about 13 km distance along the satellite track.  The instrument is not really sensitive to the very fine scale (1 km) structure that may be present in aurora. 

Some comments about auroral energy input

Over the years we have collected every observation of an energy flux of more than 60 mW m^-2 – this corresponds to a very bright aurora.  The vast majority of these instances were in the pre-midnight MLT sector – roughly between 1800 and 2400.  The probability of encountering such large energy fluxes fell by a factor of 10 or so after midnight.  These regions of very large energy fluxes tend to be pretty limited in extent, rarely more than 100 km in latitude and more often 20 km or less.   Moreover, there is the indication from ground based observations that such intense events are short-lived, perhaps 10 to 20 minutes.  The large energy fluxes are carried totally by electrons of energies from 5000 eV up to the maximum the TED measures, 20,000 eV.  This means that locally, at least, the ionosphere E-layer will be pretty disturbed but how a HF signal might be affected by such a small-scale perturbation is hard to determine.  

Things change after midnight, e.g. in the 00 to 06 MLT sector.  Here the local energy fluxes into the atmosphere tend to be much more modest (a few to perhaps 10 mW m^-2) and often 5-10% of the energy flux is carried by protons – the rest by electrons.  The particle precipitation tends to occur over a much greater extent in latitude, sometimes 10º or more, and is much more uniform.  When the energy fluxes are integrated over area, generally there is more total particle energy dump into the atmosphere in the post-midnight hours than in the pre-midnight.  The large area over which energy is being deposited more than compensates for the generally lower local energy inputs.  However, the aurora that is produced appears unimpressive to the eye, being uniform, extending over wide areas, and lacking the contrast that the eye is sensitive to.  In the post-midnight there is often an admixture of more energetic electrons in the precipitation – ones that ionize deeper in the atmosphere.  Measurements of the absorption of 30 MHz radio signals often show long lived periods of 1 or 2 dB absorption in the early morning hours that suggest degraded HF propagation through this time sector. 

How we use the observations to estimate auroral activity 

We felt that if we could estimate the total energy deposited into the polar atmosphere from these precipitating particles, we would have a good indicator of the level of auroral activity.  However, the satellites measure the energy input only along the satellite track and so we had to develop a way of jumping to a global estimate from these local measurements. 

If the energy input were uniform in longitude over the range 90º in longitude either side of the measurement we made, then it would be simple to do the arithmetic.  We could average the data over some latitude interval (we chose to average over 1-degree intervals in dipole magnetic latitude) which would correspond to some latitudinal distance in km.  One degree in magnetic latitude would be 110 km.  We could then calculate the area of a 110 km wide latitude strip, extending a distance a quarter way around the globe either side of the actual observation (taking into proper account the latitude of the observation.)  We could then multiply those areas by the averaged energy flux into the atmosphere that we observed at each latitude interval and add them up.  The data taken as the satellite ascended toward the maximum magnetic latitude would provide half the global estimate while the data taken as the satellite descended away from the maximum magnetic latitude would provide the other half of the estimate of the global energy input. 

Unfortunately, the assumption of uniform energy input over 180 degrees in longitude is a real bad one.  Because of the non-uniform precipitation and because of the nature of the POES satellite orbits (see below), this quick and dirty way of estimating global energy input gave results that had very large differences between the northern and the southern hemisphere and large variations during the course of the day (midnight to midnight, Universal Time.)  This happened even when activity was more or less constant. We had to find a way of doing better. 

We did this by creating our statistical patterns of auroral energy input – patterns that mimic to first order the spatial variations in the aurora.  These patterns were in terms of CML and MLT (not geographic coordinates.)  We did this by first using the K 3-hour magnetic activity index as a proxy for an auroral activity indicator.  We took all the available POES observations of auroral energy input, each one tagged with a CML, MLT, and K index and created our first set of global statistical patterns.  We could model the satellite orbits through these statistical patterns, and calculate at each point along the orbit the CML and MLT so that we could pluck from the pattern a statistical energy flux value.  In this way we could put together a ‘synthetic’ data set that mimicked a real satellite transit over the polar regions. 

How the nature of the POES orbits comes into play

 The POES satellite orbits are ‘sun-synchronous’, retrograde polar orbits – the inclinations are around 98º.  Sun-synchronous means that the satellite passes over a given location on the earth at two distinct local solar times – one when the satellite is northbound, the other when it is southbound.  The orbits are generally labeled by the local time, at the sub-satellite point, when the satellite crosses the equator northbound.  This equator crossing local time is almost invariant during the life of the satellite – a combination of the moon’s influence and the oblateness of the earth couple to counteract the normal precession of the satellite orbit that would otherwise occur as the earth moves around the sun.  These types of orbits are great for imaging as lighting conditions at a given location on the earth don’t change from day to day.  However, these orbits are not so good for auroral studies. 

Right now the NOAA-15 orbit crosses the equator northbound at about 1900 local, NOAA-16 at about 1400, and NOAA-17 at about 2200.  The satellites always move westward toward earlier local time.  Consider the NOAA-15 orbit.  As it moves into the Northern Hemisphere, the local time of the observation moves into the afternoon hours and the satellite reaches its maximum latitude of 82ºN at about 1300 local time (not a hot bed of auroral activity.)  It then begins its descent toward the equator moving through the late morning hours crossing 45ºN latitude southbound at about 0800, local solar time.  In short, NOAA-15 does not make an especially effective transit through the northern auroral zone.  On the other hand it does a fine job in the south, crossing the equator southbound at 0700 local time, reaching its maximum southern latitude excursion at about 0100 and then samples the pre-midnight auroral zone as it moves back toward the equator.  NOAA-16 and NOAA-17, being closer to a ‘noon-midnight’ orbit do a better job in sampling the Northern Hemisphere auroral zone. 

There are further complications because the magnetic pole is offset from the geographic pole.  There are some times during the course of a day when the satellite will go right over the magnetic pole and make measurements right on up to 90º geomagnetic latitude.   At other times the satellite will reach its maximum northern geographic latitude over on the Russian side of the Northern Hemisphere where the geomagnetic latitude is only about 70º. 

Any method of getting a better estimate of global energy input had to take all these factors into account. 

Back to synthetic data

For each of the K index statistical maps, we created a set of synthetic ‘polar passes’, determined by the geographic longitude and Universal Time of the prior equator crossing – northbound crossing for the northern hemisphere and southbound for the southern.  For each of these model orbits, we created a set of energy flux measurements using the statistical pattern and then made the ‘quick and dirty’ estimate of global energy input as described above.  However, we knew the actual global energy input from summing over the statistical pattern.  We then formed the ratio between the actual total energy input and the ‘quick and dirty’ estimate to get a correction factor – which we call a normalizing factor.  

These normalizing factors range from about 0.5 for satellite passes that transit the heart of the night time auroral zone to as high as 5 or 6 for those passes that only touch the noon time sector of the auroral zone.  For example, a NOAA-15 orbit that crosses the equator northbound at around 150º East longitude will make a very ineffective transit of the statistical auroral zone; it will not get to the highest magnetic latitudes, and the energy fluxes it will measure will generally be small.  The ‘quick and dirty’ estimate made for such a pass would be multiplied by large normalizing factor to obtain our final estimate.

We used the normalizing factors we obtained from the K index statistical maps to make total power estimates for all the satellite passes we had.  We then binned the passes into 10 estimated power intervals and created 10 new statistical patterns – now ordered by estimates of global power input.  We repeated the creation of synthetic polar passes and got a new set of normalizing factors, which we then used to create a granddaughter set of statistical patterns.  We continued this process until the new statistical patterns we obtained were essentially the same as the previous generation.  The last generation of ‘normalizing’ factors are the ones we use today.  The reason we choose 10 activity levels was because we were advised that we should try to mimic the K magnetic activity index that ranges from 0 to 9.  Folks were familiar with the 10 levels of K and might be more receptive if the auroral activity also had 10 levels. 

There is a set of ‘normalizing factors’ for each of the 10 statistical patterns, calculated for every 7.5º in geographic longitude of equator crossing, and for every 30 minutes in Universal Time.  The combination of longitude of equator crossing and Universal Time handles POES satellites in all local time orbits and for all times during the course of a day.

The actual power levels for each of the 10 bins were chosen in the following way.  We needed a minimum number of satellite passes at low power levels to get a reasonable statistical pattern.  Of the 100,000 or so satellite passes we had data for at the time, we figured about 1500 was the minimum.  There were about 1500 passes that had estimated powers less than 2.5 gW and so 2.5 gW defined the maximum power for our lowest activity level.  We did much the same thing to define the low end of the activity level 10 bin, that is chose something like the highest 1500 power estimates which turned out to be everything above about 96 gW.  We then defined the power boundaries for activity levels 2 through 9 as a geometric progression in powers – a factor of about 1.58 between the minimum power for one activity level and the next higher.  Because the activity level 10 statistical pattern was created using data from all passes greater than 96 gW and included all the data from very intense storms, the pattern looks a bit ratty compared to the other 9. 

To sum up: When the satellite completes a transit over the polar region (defined by magnetic latitudes above 45º) we use the averaged energy flux observations to compute the ‘quick and dirty’ estimate of total power input.  We know both the geographic longitude and Universal Time that the satellite crossed the equator just prior to starting this polar pass and use this information to get the 10 ‘normalizing factors’ appropriate to each of the 10 activity levels.  One final interpolation, using the ‘quick and dirty’ estimate as an input, yields a final normalization that is applied to the original ‘quick and dirty’ estimate to obtain our final (after a further minor correction for differences in sensitivities between instruments on the different satellites) estimate of global power input.  Based upon this final estimate, we assign one of the 10 activity levels and the corresponding statistical pattern of auroral energy input to that satellite transit over the polar regions.   

The final presentation of auroral activity and words of caution

As mentioned above, the statistical pattern of auroral energy input is in terms of CML and MLT– not very useful for putting on the web.  Knowing the Universal Time of the global power estimate we can compute a MLT for every 1º in geographic latitude by 2º in geographic longitude box.  Moreover, for each of those geographic boxes we can compute a CML.  So when we draw our geographic presentation (which are in terms of 1º latitude by 2º longitude pixels) we can pluck from the appropriate statistical pattern the entry for that CML-MLT location.  This is what we do for our presentation, with the 270º East longitude always at the bottom (typical American ego-centralism) and the arrow pointing to the noon meridian to provide some local time reference. 

On the main ‘auroral activity’ web page, there are two universal times given.  The first is labeled ‘Center time’ and is the time at the center of the most recent transit over the polar regions that provided the observations used to estimate the level of auroral activity.  (Almost always that transit was over the southern hemisphere.)  The second Universal time is given just above the displays showing the statistical oval plotted over the northern and southern hemisphere geographic maps.  This second time will be very close to the current time and is updated every 10 minutes.  The red arrow pointing to the noon meridian and the statistical auroral pattern are rotated clockwise the appropriate amount at the time of this update. 

The statistical pattern we plot should be viewed as only our best guess of what the global situation might be.  The averaging that went into creating the patterns ‘washed out’ all of the fine spatial structure that in reality exists in the auroral particle energy input to the atmosphere.  The boundaries of the precipitation (both on the pole-ward and equator-ward edges) have been blurred by the averaging.  The maximum fluxes in the average patterns are only about 10 mW m^-2, far less than the instances of 100 mW m^-2 or more that are occasionally observed.  Again this is a result of the averaging. 

There are versions of the auroral activity plot (accessed through the ‘Recent Data Plots’ link) where we superimpose the actual satellite data that were used to make our estimate of auroral activity on the statistical pattern.  We do this so the viewer can judge how well the statistical specification matches the actual observations, at least along the satellite track.  In general, we are happy with the agreement between the actual equator boundary of the auroral particle precipitation and the location in the statistical pattern.  They are generally within a few degrees of latitude of one another.  The agreement becomes worse at very high activity levels; estimated powers above 200 gW.  One reason for this is that the level 10 pattern was created using all passes with more than 96 gW estimated power and the results were dominated by those passes with powers in the 96-120 gW range.  The level 10 pattern often underestimates the situation during very active times when the observed equator boundaries of activity are as much as 5 and 10 degrees further equator-ward than the pattern shows.  

The agreements between data and statistics on the pole-ward boundary are generally less good.  

Another thing to watch out for are cases where we make an estimate of global power input from a satellite pass that made a very ineffective transit through the auroral region.  NOAA-15 going over the Northern Hemisphere is particularly bad in this respect (e.g. the pass on your web page for Oct. 27 at 0731 UT.)  Here the satellite sampled the 10 AM through 3 PM sector, normally a time sector where energy inputs are very low.  On this pass, the satellite sampled a location over about 2 PM with pretty high energy fluxes and, elsewhere, the energy inputs extended to unusually low latitudes (e.g. over around 10 AM.)  Based upon those measurements our approach to estimating global power jumped to the conclusion that there was a lot of energy input over the nighttime sector far from where we made the observations.  This is very likely resulted in an over estimate of the level of auroral activity that actually was occurring.  The key to determining how much confidence one should put in our estimate of global power input (and the corresponding activity level) is this ‘normalizing factor’.  The larger this number is, the less confidence one should have.  The 0731 UT NOAA-15 pass had a normalizing factor of about 4.7.  Our FTP web site (that lists every satellite pass along with its power estimate) has all the normalizing factors.  Sometime beginning in mid-January or so we will include the normalizing factor on the auroral activity page along with a sentence explaining its significance. 

A few words about how the data arrives in Boulder, CO

You may have noticed that usually the most recent estimate of auroral activity came from a Southern Hemisphere satellite pass.  This bias is connected with the way NOAA gets the data from the satellite and passes the data to us in Boulder. 

The satellites are operated on a data store and dump basis.  Data dumps are made to stations in Fairbanks, AK and Wallops Island, VA when the satellites are above the radio horizon.  Normally one orbit of data is transmitted down and when the playback is initiated, the on-board data stream is switched to another recorder that will not be read out for one orbit or more in the future.  Once a playback is completed, the data are relayed to a facility in Washington DC that strips out the SEM data and sends a file to us.  The time between the end of a playback to a station and when we get the data may run 20-30 minutes.  

The problem we face is that we require data from a full transit over the polar region before we make the estimate of auroral power and level of activity.  When the playback is made to Fairbanks, only the first part of the Northern Hemisphere transit is present on the recorder being played back – the second half is now being recorded and will not be available until the next ground station contact.  Thus the most recent complete polar transit that Fairbanks gets is the last Southern Hemisphere transit.  Contacts to Wallops Island are a little better.  If the satellite is southbound when Wallops Island makes the contact, the satellite may have already exited the northern auroral zone when the playback starts and, so, we will get a complete Northern Hemisphere transit in a very timely fashion.  However, the bulk of the playbacks from these polar satellites are made through Fairbanks and not Wallops Island. 

There are plans in the future that will go a long ways to mitigate this problem. Plans (8-years from now) call for a network of ground stations so that these satellites are rarely out of radio contact.  The objective is to get data to the user no more than 15-minutes after the satellite has acquired it.  When that happens, the estimates of auroral activity will be much timelier. 

One thing that we have talked about doing now is to identify the location of the equator boundary of the auroral precipitation.  That location is an indicator of activity – the lower the latitude, the higher the activity.  The data dump to Fairbanks would contain the equator location for the ascending phase of the Northern Hemisphere pass.  That would give an idea of activity level nearly 45-minutes later than the time-tag associated with the completed Southern Hemisphere pass.

I might also mention in passing that the data from the SEM are transmitted down from the satellite continually in the 136 MHz band along with the cloud imagery.  There are people around the world that capture this signal to produce cloud images in real time as the satellite goes over.  They invariably throw away the bits that contain the SEM data.  I have only heard rumors that the Japanese in Antarctica and, perhaps, the Danes in Greenland extract the SEM data from that down link to get real time observations of the particle precipitation in their vicinity.  If anyone is interested in extracting the SEM data, I can help with formats and ways of converting the transmitted 8-bit words to physical quantities. 

The more energetic particle instruments 

Earlier I mentioned that there were additional instruments in the SEM that monitored charged particles of much higher energy than the auroral instrument.  Specifically we monitor the fluxes of electrons of energies greater than 30,000 eV (30 keV), greater than 100,000 eV, and greater than 300,000 eV.  While 300,000 eV electrons get down to 70 km in altitude before they are stopped, the fluxes are normally so low that there is little effect on D-layer electron densities.  The fluxes of greater than 30,000 eV electrons can be significant, especially in the early morning hours, and will contribute to ionization in the lower E-layer and HF absorption.  However, to be honest, we have never integrated the 30,000 eV electron observations with the 50 eV-20,000 eV auroral particle observations in any of our presentations. 

The instruments also monitor proton fluxes over a very wide energy range from 30,000 eV all the way up to solar energetic particles of energies 140,000,000 eV (140 meV).  The fluxes of protons of energies below 800,000 eV or so don’t contribute much to disturbing the ionosphere except during large magnetic storms.  During the large storms, the energy fluxes of >30,000 eV protons may get up to 10 mW m^-2, a level that would produce significant ionization.  However, at the same time during these large storms many other processes are disturbing the ionosphere and the contribution of the proton influxes is hard to untangle from these other processes.  As with the energetic electron observations, we have not integrated the energetic proton observations with the auroral particle data in any of our presentations. 

There is a set of four solid-state detectors in the SEM that monitor energetic solar protons during solar particle events.  The guys are quite energetic, above 10 meV.  The GOES satellite monitors the intensity of these particles out at geostationary orbit.  The POES satellites monitor the fluxes at low altitudes and provide information about the spatial extent over the polar regions impacted by these particles entering the atmosphere.  The polar region D-layer is really enhanced during solar particle events – the incoming protons are energetic enough to get down to 50-60 km (and sometimes even lower) and produce a great deal of ionization as they come to a stop in the atmosphere.  As you know, ionization in the D-layer is especially effective at absorbing HF radio signals.  During solar proton events HF propagation across the polar regions may be completely impossible (polar cap blackouts.) 

We do have presentations that provide some information about these very energetic particle fluxes, but I think they are somewhat more specialized and less useful to the amateur radio community (except, perhaps, for the web page on solar protons.) 

Summary References

In closing, let me list some of the POES/SEM SEC web pages that you might find useful along with a brief description of their content. 

www.sec.noaa.gov/pmap     This is most recent estimate of the auroral activity.  Once the activity level has been determined, the statistical pattern is plotted for both hemisphere irrespective of which hemisphere the actual observations were made over.  If you click on one or the other image, you get a blown up version. 

www.sec.noaa.gov/pmap/Plots.html      This brings up the list of most recent individual satellite passes.  Clicking on an entry will display the appropriate statistical pattern along with the energy flux observations made by the satellite which can aid in interpretation. 

www.sec.noaa.gov/ftpdir/lists/hpi/pwr_1day   This brings up the current day’s list of estimated powers with the corresponding ‘normalizing factors’ that indicate the confidence in the power estimate. 

www.sec.noaa.gov/ftpdir/lists/hpi/pwr_1day.bak  This brings up yesterday’s list of estimated powers with the corresponding ‘normalizing factors’ that indicate the confidence in the power estimate. 

www.sec.noaa.gov/ftpdir/lists/hpi/     This brings up a menu of all historic estimates of power input ordered by year.  The ordering in each year’s list is by day of year, not by date, and that may be an inconvenience. 

www.sec.noaa.gov/ftpdir/lists/hpi/plots/     This link provides access to the previous 7-days of auroral activity plots – NOAA-15 data only but we are thinking of expanding it to contain NOAA-16 and NOAA-17 data as well. 

http://solar.sec.noaa.gov/~sgreer/view/view_any_gif.cgi   This web page is the one we use to satisfy external requests for specific plots, for example from periods of very high activity.  There are currently pmap plots for March 31, 2001 that you may find of interest.  There are also some plots of >30 keV electron and proton precipitation during the April, 2002 storm period as well as some plots of solar proton influxes from the April 21-23, 2002 event.  The entries on the menu might be a bit confusing but as a rule the ordering is year – month – day – UT with the text indicating the content of the plot.  One highlights the entry of interest and then click on submit to get the plot. 

www.sec.noaa.gov/tiger/index.html    This web page contains a display of the most recent >30 keV electron intensities – perhaps not in a very useful form.  The color code is the departure of the observed intensities from their one-year average values.  Red means the observed intensities are 10 times higher than the one-year average at that location.  This form of display is intended to provide a quick idea of whether, and where, the particle intensities are unusually high.  However, the display really does not provide much quantitative information about how disturbed the ionosphere might be.  The link to ‘Recent Data Plots’ gives access to similar plots for all the particle energy channels. 

www.sec.noaa.gov/tiger/day1html/30-80keVp90.html  This web page is similar to the >30 keV electron one but is for >30 keV protons.  During large magnetic storms the proton intensities go way up.  When Kp is high, you might look at this page and see where all the red is. 

There are two displays that provide information during solar particle events.  They are:

www.sec.noaa.gov/tiger/day1html/16-70MeVomnip.html    and

www.sec.noaa.gov/tiger/SolarProtons.html

The presentations are somewhat different but both give a good idea of the extent of the spatial disturbed D-layer during solar particle events and those HF propagation paths that will be badly degraded. 

And, of course, there is the web site that shows the solar particle measurements made by GOES.  These data have the advantage of being in real time.

www.sec.noaa.gov/today.html   or    www.sec.noaa.gov/rt_plots/pro_3d.htm