Loading ...
Sorry, an error occurred while loading the content.

Re: [dxatlas] 90-day SSN data for HamCAP

Expand Messages
  • Alex, VE3NEA
    A while ago I did some comparisons of median SNR predictions from VOACAP to the observations of NCSXF beacons. Ingemar SM5AJV developed a program that measures
    Message 1 of 17 , Apr 9 10:10 AM
    • 0 Attachment
      A while ago I did some comparisons of median SNR predictions from VOACAP to
      the observations of NCSXF beacons. Ingemar SM5AJV developed a program that
      measures the SNR of beacon signals using a sound card and an FFT
      transformation. A correlator is used to compare the received waveform to
      that of the beacon's callsign, to make sure that the signal actually comes
      from the beacon. Ingemar was kind enough to send me his logs for June 2004,
      these logs cover about 24% of the month.

      I developed a viewer utility for the logs that converts the observed SNR
      values
      into dB-Hz, calculates medians, and plots both predicted and observed median
      values along with the observation points for the selected beacon and band.
      The program lacks a sophisticated user interface, which I may develop in the
      future if there is sufficient interest. A few screenshots of typical plots
      are here: http://www.dxatlas.com/Private/SnrPlot.zip .

      For those who want to play with the viewer, I uploaded it to
      http://www.dxatlas.com/Private/VoaAnal.zip .
      The zip file also includes the observation data, with permission
      from Ingemar. Please preserve the directory structure when unzipping.

      Though this is just the first alpha version of the program and thus it may
      contain errors, the plots it produces are very interesting. The shape of the
      VOACAP prediction chart is very close to that of the observation chart,
      though the absolute values are way off. On average, VOACAP underestimates
      the SNR by 20+ dB.

      Also, these charts explain why the median SNR is a bad indicator of
      propagation conditions. The median value makes sense only for unimodal
      distributions, while the distribution of SNR is bi-modal at best (that is,
      has two or more peaks). One peak is formed by the observations that where
      performed when the path was open with a good propagation mode, and one or
      more peaks come from the days when only very lossy modes were available or
      the path was closed.


      73 Alex VE3NEA
    • Pete Smith
      OK - thanks! Makes sense. 73, Pete
      Message 2 of 17 , Apr 9 11:27 AM
      • 0 Attachment
        OK - thanks! Makes sense.

        73, Pete

        At 12:56 PM 4/9/2005, Alex, VE3NEA wrote:



        > > While we're on the subject, a question I have wondered about -- why do Ham
        > > Cap's charts not have a graduated hour scale on the X axis? Have I missed
        > > an option to turn them on?
        >
        >HamCap is an interactive application. When you move the mouse cursor over
        >the chart, the time and all other parameters of the point under the cursor
        >are displayed on the status bar.
        >
        >I tried to make the HamCap window as small as possible so that it could be
        >used during contesting and DXing when screen space is a valuable resource,
        >so I removed all unnecessary elements form the user interface, including the
        >time scale on the chart.
        >
        >
        >73 Alex VE3NEA
        >
        >
        >
        >
        >
        >
        >Yahoo! Groups Links
        >
        >
        >
        >
      • Pete Smith
        [With apologies to readers of the VOACAP list, who are joining this in mid-thread, I feel that Bill and I are floundering fairly far out of our depth in this
        Message 3 of 17 , Apr 9 11:32 AM
        • 0 Attachment
          [With apologies to readers of the VOACAP list, who are joining this in
          mid-thread, I feel that Bill and I are floundering fairly far out of our
          depth in this discussion that began on the dxatlas list. The question
          began as what SSN to use in Ham Cap, a simplified front end for
          VOACAP. HamCAP comes with a table of predicted international SSNs, but can
          also link with IonoProbe, which downloads daily SSNs from the SEC. VE3NEA,
          the author of Ham Cap, prefers the SEC numbers, but the variance from the
          international SSns is considerable, with the SEC number typically running
          much higher. Any advice appreciated!]

          Bill, there's some apples and oranges, or something going on
          here. IonoProbe gets its numbers from the Joint USAF/NOAA Solar and
          Geophysical Activity Summary. Here's a sample:

          "SGAS Number 099 Issued at 0245Z on 09 Apr 2005
          This report is compiled from data received at SWO on 08 Apr
          A. Energetic Events
          Begin Max End Rgn Loc Xray Op 245MHz 10cm Sweep
          None
          B. Proton Events: None
          C. Geomagnetic Activity Summary: The geomagnetic field was quiet.
          D. Stratwarm: Not Available
          E. Daily Indices: (real-time preliminary/estimated values)
          10 cm 088 SSN 056 Afr/Ap 005/004 ...."

          Ionoprobe's retrospective record of these reports shows March 12 (chosen
          randomly) with a SSN of 67, versus 42 on the table you cite. March 18 is
          37 versus 25, and so on.

          I have no idea what the explanation is, though I do note that there is
          another table, ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/2005
          , which gives the numbers just for this year to date and states that
          "Values are preliminary after Dec 2004."

          There is an interesting explanation of how various sunspot numbers are
          calculated in
          ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/sunspot.predict. It
          seems clear from this explanation that the international sunspot number is
          a highly-massaged, worldwide average number that does not become final for
          some months after the date of observation. By contrast, the number cited
          in the SGAS is the SEC's every-six-hour figure, and must be based strictly
          on its own observations. An interesting graph at
          http://www.nwra-az.com/spawx/comp.html speaks to this. To quote from the
          caption:

          " This plot illustrates the differences between the "real" sunspot number
          (SSN), which is calculated from optical observations of the sun, a sunspot
          number derived from the 10.7cm solar radio flux (SSNf), and a sunspot
          number derived from fitting an ionospheric model to ionospheric
          measurements. All of these indices are used as inputs to models of the
          ionosphere for use in communications-performance predictions - this plot
          shows that they don't always agree as to what the SSN should be in that
          particular context.

          Note: The F10.7-derived SSN (SSNf) is calculated from the 10.7cm solar
          radio flux (the Penticton Radio Observatory noon value) using the following
          relationship:

          F10.7 = 63.74 + 0.727*SSNf + 0.000895*SSNf**2

          So, where does this leave us? Where it left me, frankly, is wondering why
          we use sunspot numbers at all, rather than solar flux and A/K indices, or
          at least the SSNf, which would be closer to the values Bill cited than to
          the current high value of optically-observed SSN. On the other hand, Alex
          argues that results from using the lower SSNs seem to run consistently low
          in terms of predicted vs. observed S/R ratio.

          It would be interesting to pass this discussion over to the VOACAP list and
          see what the gurus there, particularly including Greg Hand and George Lane,
          think of it. In fact, I have done that, and it will be interesting to see
          what comes of it.

          73, Pete N4ZR



          At 09:12 AM 4/9/2005, bill_w4zv wrote:



          >I wrote:
          >
          > > Maybe the logic of
          > > the calculation has an error somewhere in its data collection.
          >
          >Definitely so. The current 90 day average (1 Jan 05 - 31 Mar 05) is
          >28.4 using this data:
          >
          >ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/RIDAILY.PLT
          >
          >I don't have IonoProbe, but it must have a problem if the 47 number
          >Pete quoted for the IonoProbe calculation is correct. Anyhow, now I
          >know what to use and will probably just go with the NOAA forecast
          >which seems to be fairly accurate at this stage of the cycle.
          >
          >73, Bill
          >
          >
          >
          >
          >
          >
          >Yahoo! Groups Links
          >
          >
          >
          >
        • bill_w4zv
          ... wondering why ... I agree. Seems like measured solar flux, which supposedly shows the actual effect of sunspots on the ionosphere, would be better. After
          Message 4 of 17 , Apr 10 4:40 AM
          • 0 Attachment
            --- In dxatlas_group@yahoogroups.com, Pete Smith <n4zr@c...> wrote:

            >
            > So, where does this leave us? Where it left me, frankly, is
            wondering why
            > we use sunspot numbers at all, rather than solar flux and A/K indices,

            I agree. Seems like measured solar flux, which supposedly shows the
            actual effect of sunspots on the ionosphere, would be better. After
            all, when there is a large CME, we often can see little effect on the
            ionosphere depending on how it's directed toward Earth, which way Bz
            points, etc. Of course if VOACAP was originally based on SSN's, then
            maybe they should be the input. Hopefully Greg Hand will respond, and
            please post anything of interest here since I am not on that list. No
            model is any better than its input..."Garbage In Garbage Out" as
            someone said.

            73, Bill
          • bill_w4zv
            I wrote: The current 90 day average (1 Jan 05 - 31 Mar 05) is 28.4 using this data: ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/RIDAILY.PLT VE3NEA
            Message 5 of 17 , Apr 10 12:50 PM
            • 0 Attachment
              I wrote: The current 90 day average (1 Jan 05 - 31 Mar 05) is
              28.4 using this data:
              ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/RIDAILY.PLT

              VE3NEA wrote: These are the RI indices, not daily SSN values.

              Alex, the NOAA site below calls these "Provisional International
              Sunspot Numbers, so what do you mean they are not daily SSN numbers?
              See especially their comments preceded by ***** below:
              ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/sunspot.predict

              >The provisional daily Zurich relative sunspot numbers, Rz, were based
              upon
              observations made at Zurich and its two branch stations in Arosa and
              Locarno
              and communicated by M. Waldmeier of the Swiss Federal Observatory.
              Beginning
              January 1, 1981, the Zurich relative sunspot number program is
              replaced by
              the "Sunspot Index Data Center" (c/o Dr. P. Cugnon, 3 av. Circulaire,
              B-1180
              Bruxelles, Belgium).


              ***** The determination of the provisional International Sunspot
              Numbers Ri results from a statistical treatment of the data
              originating from more than twenty-five observing stations. These
              stations constitute an international network, with the Locarno
              (Switzerland) station as the reference station, to guarantee
              continuity with the past Zurich series of Rz.*****

              >The definitive International Sunspot Numbers Ri are evaluated by a
              similar method based on a network of observing stations selected for
              their
              high number of observations, their continuity during the whole year
              and an
              existing series of observations during the last years. Also taken into
              account is the stability of the K monthly factors with reference to the
              Locarno station.

              *****These relative sunspot numbers are now designated Ri
              (International) instead of Rz (Zurich).*****

              It seems to me Ri is exactly what should be used, but am really
              waiting to see what Greg Hand, etc. of VOACAP have to say.

              73, Bill
            Your message has been successfully submitted and would be delivered to recipients shortly.