I believe that’s page 5 of this
Although TBH, I think attenuation in vegetation is what happens between the front & back of a GCSE Computer Science class when I’m live editing a GIF at hex level.
And the conclusion is: Using LoRaWAN it is rather unlikely to cover more than 1500 m in a forest.
But this could be a field for more experiments and measurements.
eerrmmm! see comments/experience above!
Hello, I have been working on water level measurement (Australian water tanks) for about a year. And now I have some good results. I am working in the South America Chaco, so I have a lot of forest. The farest tank is about 30km from the Gateway (I am using a RAK Bundle Pro). I had been using Dragino Shield with arduino and LMIC, but now I am using a hardware design that I made. with a small solar panel and a 18650 battery. If there is gonna be a long distance between your electronic and sensor it is better to use a 4-20mA sensor in order to compensate the distance but is power hungry. I have done this way in the first attemp. Putting my sensor in the tank and about 40m away the electronics with some gain of altitude. Now I am using an ultrasonic sensor that is 20cm from the electronics but it is working. You can use https://www.ve2dbe.com/rmonline_s.asp to calculate the fresnel zone with GPS position.
2 Likes
I can tell you with certainty that this should be no issue as long as both your antennas are a few meters (at least 3m) above ground. We are doing exactly what you are trying to do routinely with our farmers.
maybe i could attach the antenna in a tree
Yes, put your antenna up the tree. 
PM me if you need a sensor for the soil moisture or an application to go with it.
1 Like
Jeff, I am looking for any data that depicts LoRA coverage in scrub pine forest. Do you know of any studies I can look at? Also have you published any of your tests, like the one you show here?
Thank you,
Dolores
There are several practical examples mentioned in this thread.
One of the significant issues with predicting ‘Coverage in a Forest’, is the very nature of forests, they vary significantly in height, density and type of trees, The amount and type of ground based vegitation has a significant affect as does the underlying terrain.
The best way of predicting range is to carry out simple practical range tests at the location involved, that would give you far more accurate predictions of range.
If you describe the actual application, and where it is, you might get some good suggestions …
IIRC There are academic and use case studies posted on the net (GIYF) that may help but in the context of TTN the 3 or 4 posts I did back in Oct '20 is about the limit of what I can disclose for commercial reasons - all my efforts and deployments wrt TTN are ‘free’ but represent a significant investment in time, capital and IP where I need to protect info for consulting and customer chargeable use case benefit, so not openly published. I also have early LoRa vs specifically LoRaWAN (its the RF that counts afterall for the most part!) examples and use case info from 2013/14/15 (essentially before LoRaWAN, and specifically TTN became a thing!) from other areas including forests in Southern Finland, Sweden & California - done for lone/remote worker/equipment, forestry management and early fire detection monitoring that essentially re-enforces the key messages and advice from me and others above. Key points that influence achievable coverage:
-
Topology - even gentle undulations of the land can end up masking GW/node communication especially for low level GW’s without realising/being obvious ‘on the ground’
-
Planted vs Wilded/ancient wooded land - affects RF absorber density and viability/availability of pathways through forest
-
Managed vs unmanaged - if underbrush regularly cleared and maintained this can have huge benefit vs coverage decay that creaps in over years of neglect/wilding - also impacts abillity to manage/control wild fire outbreaks
This is especially the case for low level deployments (say GW<5-10m above local GL) -
Type and mix of arboreal growth - tree type! as well as age mix… note for planted sites coverage can be great for years then suddenly over just a couple of years you find range starts to collapse as the trees grow at roughly same rate and thickness and RF pathways start to see greater overlap and effective density. Plan ahead for increase GW densification to mitigate!
-
Seasonality and moisture content & coverage - snow/ice even rain covered affects 2.4GHz LoRa more than the 900Mhz bands but still has impact, heat stressed growth has lower impact due to lower internal water content, though some ‘sappy’/sticky trees can apparantly get worse.
Best tip (as is usually case with LoRaWAN deployments generally) is get the GWs or at least the ants high - use towers/portable masts, high slopes or even deploy to top of trees to save costs! (and height tracks vertical growth! ) Use at or above canopy level is you can vs down deep and close to ground level - same goes for the nodes, though often the nodes are needed at or close to ground level for their use case/purpose.
Sadly, I dont have much info or 3rd party knowledge on effectiveness for use in high density/wet tropical rain forest use - though a friend and former colleague was working on a specific animal tracking project for WWF on Jungle fringes a few years back - 
must check what data he gathered…
1 Like