Mon, Apr 13 at 4:48 AM
All,
North Texas Microwave Society has a team that has been at work full time in this area for 2 months. Our algorithms have changed 3 times, and we are rolling in several new aspects of analysis, data source and compute this week.

This past Saturday we presented the current state of the work at the April NTMS meeting. You can view the recording on YouTube. The propagation tools discussion begins at 47 minutes, 40 seconds. NTMS April 11 2026 meeting



There are several additional presentations planned in May/June/July and at Microwave Update 2026 Conference in Rochester, N.Y. on October 23-24  Microwave Update Presentation and Proceedings title: “Empirically Calibrated Propagation Forecasting for SHF/EHF Amateur Paths: 10–241 GHz"


Our current model incorporates a QSO dataset (1992-2026) of approximately 60,000 contacts, 3000 ASOS integrated (Automated Surface Observation System), 346 RAOB integrated (Radiosonde Observations), AWS S3 HRRR integrated (NOAA High-Resolution Rapid Refresh), 3,675 hourly IEMRE gridded observations (resolution 0.125 degrees or approx 14 km, SRTM integrated (Shuttle Radar Topography Mission), 7 commercial links validation observations (real time 11//24//65 GHz) from Ubiquiti airFiber and airFiber 60 links (need more of these across CONUS), Solar Indices 1998-2026.

Our tools include real time forecasting, specific grid square to grid square (10 digit) link closure probability in real time and future forecast up to T+12 hours. We have a CONUS contact map of all contacts in the dataset with ability to drill down to the detail atmospheric method that enabled the contact. There are additional features too numerous to mention here (watch the presentation video).

As to Kevin's ask. This is well within the range of technology on hand.  However, a 4-digit grid (XXnn) is a large area that would require significant compute to analyze. The area first must be gridded at a small enough resolution to scan for vegetation and walk the inspection out in the direction of the path required (viewshed). SRTM data (think tree/foliage canopy detection) versus NED National Elevation Data (think bare earth) could be leveraged, and these are free. GIS software, also free, can run viewshed calcs which would determine the "open paths/bearings" from the bare earth locations identified. We could blend Global Forest Watch data (NASAs GEDI lidar mission to detect canopy height - 2019 to current) also free and should be compared to SRTM to pick the most current/accurate. Trees/foliage grows and also is cleared. Some SRTM data was collected in leaf-off periods.

NTMS has a different project in queue that has a higher priority. This is a feature that will take two 10-digit grids and monitor the current and forecasted link closure probability by frequency of interest and notify subscribers with alerts. The analysis is pertinent to the geographical area represented by the path between the grids anywhere in CONUS and southern Canada. There are big differences between analyzing a path between two mountain tops in the Western U.S., an over water path in the Great Lakes area, a flat prairie land path in the Midwest and a path in the arid desert of the southwest.

The current priority task would work complimentary to Kevin's idea. Find the possible paths, then monitor for probability of a completed contact and keep your finger on the "roll trucks" button. For us in DFW, we are interested in providing efficient means for our local members (500-mile radius) to complete VUCC on 10/24/47 GHz. We have many dozens of vetted locations (free of vegetation, good clear horizon) but these paths are shutdown except when enhancement occurs. What we develop on the proving grounds here should work anywhere. 

Our learning model needs more QSO inputs and we have added a means to enter contacts data, either one at a time, CSV import or ADIF import. Please consider adding your QSOs to the dataset. Since the dataset already contains all contacts for 6 years of the 10 GHz and up contest, we are looking for contacts made outside of the contest. We are collecting contacts made on 902 MHz and up. You can also use a search feature to see what contacts the dataset has.

73,
Jim McMasters
North Texas Microwave President



On Sunday, April 12, 2026 at 11:28:32 PM CDT, J. Gordon Beattie, Jr. W2TTT via groups.io <[email protected]> wrote:


Kevin VE3KH,
OK... to quote Curtis Sliwa, "Sleep is for the dead."  😀

Here is what I posed to CLAUDE.AI

I would like to give two locations in 6 character maidenhead grid square format and present the geographic contour between the two locations

I got this response.

I would like to give two locations in 6 character maidenhead grid square format and present the geographic contour between the two locations

11:34 PM

It took 24 minutes and another 2 minutes to test it, but here's what I was able to produce.  


This first one used grid squares that the tool selected.  

The second one is for a local path between my shack at W2TTT and the  K4DBC shack.

Image

Image

It took several iterations to get by some errors, but ultimately I was able to hash this out.  
There was a note at the bottom that read:
"Real elevation data via Open-Meteo (SRTM/Copernicus 90 m). Path follows linear interpolation between grid square centres."

This is pretty neat stuff.  It ultimately produced two methods and only one works (for now).  I would like to get some sleep before a full work day and work week.


More later!

73,
Gordon Beattie, W2TTT 
201.314.6964 

Get On The Air!

From: [email protected] <[email protected]> on behalf of J. Gordon Beattie, Jr. W2TTT via groups.io <[email protected]>
Sent: Sunday, April 12, 2026 11:32:13 PM
To: [email protected] <[email protected]>
Subject: Re: [10GHz-Up] Grid Contact Best Path Profile
 
Kevin VE3KH,
I made a suggestion about how to use CLAUDE AI to build out the path prediction tooling that you were seeking.

We just need someone with the time to play with it.

73,
Gordon Beattie, W2TTT 
201.314.6964 

Get On The Air!

From: [email protected] <[email protected]> on behalf of VE3KH FN03CG Kevin via groups.io <[email protected]>
Sent: Sunday, April 12, 2026 11:00:49 PM
To: [email protected] <[email protected]>
Subject: Re: [10GHz-Up] Grid Contact Best Path Profile
 

I really thought that with all the AI and Chat GPTs and so on there might be a chance to improve the tools we use.

Kind of like what Rainscatter did for 10GHz !

 

 

From: [email protected] <[email protected]> On Behalf Of K3WGR Allen FN10XI via groups.io
Sent: April 12, 2026 8:46 PM
To: [email protected]
Subject: Re: [10GHz-Up] Grid Contact Best Path Profile

 

YEARS  ago when roving we used Google Maps and then google street view to hunt for roving spots that were high and also clear site - the ability of google maps to zoom in and rotate around site really help to either accept the site as workable or perhaps only open to one direction that may or may not be open to possible Q’s - you can also look for parking spots.  This also helps to look for sites in grids that are not heavily populated (rare or little used grids)

 

Allen K3WGR  rover van NN3Q/r

 

73 to all



On Apr 12, 2026, at 6:06 PM, VE3KH FN03CG Kevin via groups.io <[email protected]> wrote:

 

Always try

But planning is always a good start 

 

Thanks, KH



On Apr 12, 2026, at 5:41 PM, N8IUP Michael EN81RJ via groups.io <[email protected]> wrote:



Ya, but ya gota try!!!! 

 

On Sunday 04/12/2026 at 1:52 pm, VE3KH FN03CG Kevin via groups.io wrote: 

I agree … Webmap works well … but identifying the highest points or best path to an area is manual and you may never stumble onto the best path.

 

 

 

From: [email protected] <[email protected]> On Behalf Of K1FMS Fred FN32QB via groups.io
Sent: April 12, 2026 1:42 PM
To: 10ghz and Up Groups.IO <[email protected]>
Subject: Re: [10GHz-Up] Grid Contact Best Path Profile

 

I use a combination of K7FRY and the W3SZ / K1RZ database map.  Works well with 10 digit grids.

73

Fred

K1FMS

 

On Sun, Apr 12, 2026, 1:37 PM VE3KH FN03CG Kevin via groups.io <ve3kh=[email protected]> wrote:

To all the software guru’s … 

 

It would be nice to have a tool that works with Hey Whats That (or similar) to find the best place for a portable to setup to work based on the best path profile.

 

I am thinking something like:

·    Enter your 10 digit grid

·    Enter the desired location you are trying to contact or an area where another station could setup  … the number of grid characters would specify the size of the target area.

·    Up pops the best Path Profile

 

Rought example:

·    Say I want to work EN93 as a new grid

o  I Enter my location accurately … FN03CG58LO

o  I want to work EN93, and its really convenient to get to EN93XI (close and good roads)

§ The APP figures out where the highest point in EN93XI is and draws the path profile between the 2 places

 

Anyone have any ideas?

 

73 Kevin VE3KH

 

 

 

 

_._,_._,_
Groups.io Links:

You receive all messages sent to this group.

View/Reply Online (#10461) | Reply to Group | Reply to Sender | Mute This Topic | New Topic
Your Subscription | Contact Group Owner | Unsubscribe [[email protected]]







_._,_._,_