GOOD SITE SURVEYS
IN ALASKA
A high
view of the Tanana River
With the request by the Biological Sciences
Division, NSF, that we add the Bonanza Creek, Alaska, LTER to
our wireless test modeling project, and the approval of a budget
supplement to our basic project funding to cover the added costs,
we commenced the first project work August 27-31, 2000. Much later
and we would have faced the legendary winter weather of central
Alaska.
Mike Willett, technical assistant, and myself,
as PI, rendezvoused in Fairbanks the 27th. The Bonanza Creek LTER
Project had just added Wendy Davis, from Canada, as the LTER Manager.
So we had a single point of contact to work with for both the
Bonanza Creek Experimental Forest Area, 16 miles down the Tanana
River from Fairbanks, and the Poker Flat Watershed area 35 miles
northeast. She did valuable advanced work for us assembling both
USGS maps of both research areas for us, and getting computerized
path analyses done and printed out which narrowed the possible
radio relay sites we could use to link to both the University
and the largest possible segment of the Tanana River where most
of the Data Loggers were deployed. We knew from my earlier reconnaissance
in June, that relays would be necessary.
Wendy accompanied us on Sunday, the 26th, for
a rough 'drive about' survey in a 4 wheel drive rental vehicle
of possible relay points we and experienced personnel from the
University felt might work, north of the Tanana River astride
Alaska Route 3 to further narrow the choices before we did radio
link tests on Monday.
RELAY POINTS
While it appeared that the highest hill in
the area, Ester Dome, which had a large number of radio, television,
and telephone facilities on it was the obvious choice, we already,
by my previous eyeball reconnaissance, saw that a range of hills
between it and the Tanana River masked large segments of the river
basin from view from that peak. With 'Line of Sight' radios such
as the FreeWave 915Mhz series we thought would be best to span
the 5 to 20 mile reaches, that would not do. A second visit to
that hill confirmed our view it was not a good location to see
the river basin from, even though it was ideal to see the University
buildings from.
So we looked more closely at several other
points, including a hill on Chena Ridge, 6 miles, and possibly
line of sight from the roof of the Arctic Health Center at the
University, within which were the offices and laboratories of
the LTER.
As it turned out this hill was the best site
of all. The three questions were (a) can we get permission to
use the peak (b) could we get a good radio link. (c) whether or
not the Fresnel Zone phenomenon - the elliptical envelope that
the radio waves travel within, requiring certain numbers of feet
of 'ground clearance' below the actual direct point to point visual
line of sight. Although the map path data showed an unobstructed
line between that hill top and the University site, another ridge
about halfway between the two points came within yards of obstructing
that sight line. And a binocular scan from ground level of the
hill, did not reveal the low dark stone building which was our
target, but partway up the white walled Museum building a few
hundred meters further east. Pictured
below is the building we want to reach.
Arctic Health
Center Building inside of which is the LTER office that we want
to reach.
To reach the peak we drove up a dead-end road,
asked a local whether we were close to the peak, and hiked up
a steep tree cut used by power poles to reach the top. Where we
found a perimeter fence that we circled seeing that several points
outside the fence line could be used to 'see' both back to Fairbanks,
and down the River. It might take a 10 or 20 foot mast to clear
the near trees, but it looked promising, as we were able to glimpse
through openings in the trees, both the University and portions
of the Tanana River where, according to both our USGS map, and
a computer path analyses the data loggers should reside.
The FAA facility
looks like this up there.
A look through the trees near
the top, which we would have to get above, shows what we see
of the Tanana.
We also drove to three other points, from
which we looked toward the river basin. One was the current LTER
#1, and fairly close to it was the abandoned - except for a 30-40
foot tower - the old LTER #1 site. Mike climbed part way up the
tower but could not see much more of the valley over the close-in
trees.
And we were able to drive close to LTER #2
site.
LTER #2 site, which is a cleared area with
a number of instruments, a small building, and one mast antenna
of about 30 feet, which has a cell phone that intermittently works.
It marks the furthest distance from the University that cell phones
operate down river. But because it was on the 'near' shore, and
the closer one gets to the ridges on the north side of the river,
the more trees the signal has to penetrate from the Chena Ridge.
So the radio link was intermittent and questionable, without a
tall mast antenna.
Busy top of
Ester Dome
We drove to the top of Ester Dome, which is
both jam packed with radio and television and other communications
towers - which could cause interference problems if any of them
were in or near the 902-928Mhz range. In particular, 'dirty' Paging
Towers, whose systems operate in the 929Mhz frequencies, cause
some problems with the spread spectrum radios -causing a slow
down in rates. But even though the Dome appears to dominate the
region, ridges and hills closer to the river block the view to
the boundaries of the original Bonanza Creek Experimental Forest
tract within which most data loggers are placed. It is interesting
that the Experimental Forest designation goes back many years,
for it appears on 1972 USGS maps sheets. So the Dome is less useful
for our purposes than those who don't practice visualizing 'sight
lines' for a living do.
We filed away in the backs of our minds the
areas we could see from Ester Dome - with the possibility it might
have to become a secondary relay site.
MONDAY - AT
THE BASE SITE
Monday, the 28th, we first scouted out the
best pathway from the top of the roof at the Arctic Health building
down inside the building to the room where the Data Logger data
was routinely downloaded from field units. This seems like a minor
matter, but often overlooked by those contemplating the use of
data radios, the matter of getting the data flow from the radios,
or antennas, or both, which are high in or on buildings, to the
room where computers can be connected to them is just as much
a matter for good planning and site survey as the long outside
radio links. RS232, or Ethernet cables can be strung between radios
and their destinations. But running even the best quality RF cables
between a radio deep inside a building, and then to a roof antenna,
guarantees 'line loss' that can approach 4 dB per hundred feet.
Which attenuation of the signal can make long radio links marginal
or unusable. More than 50 feet of RF cable is undesirable, even
if necessary.
As a principle, it is better to keep the RF
cable length as short as possible, even to the point of placing
the radio as close as possible to the outside antenna, even if
it is in an inside space below the roof, while running RS232 serial,
or Ethernet cable, depending on what the radio uses, to the room
where the data line connects to a computer or network. Or where
that is even difficult, or because of the construction of the
building, too costly, to use another radio at the computer side,
which talks to the radio on or near the roof, which acts as a
relay. That of course, besides cost, introduces other problems,
for many devices and software packages have trouble with multiple-hop
radio links. (We have yet to test PCW208 Campbell software and
data loggers under these conditions. We will.)
We found that the 2d story roof of the west
end of the Arctic building was the most likely place we could
put the base radio antenna, putting a small tower on it if necessary
to get height and better clear that middle ridge. And we saw that
there were no power outlets on the roof. In fact we found no open
'conduits' to the levels below, either. We were reminded that
this is cold country, and the buildings are tightly sealed. However
we saw the possibility of running a thin cable down through a
trapdoor that opened to a stairwell. And that there was a power
outlet within about 6 feet of the trapdoor.
Mike
Willett on the Arctic Health Center roof.
Now it was just a question of whether an antenna
near that location could link with a radio and antenna on Chena
Ridge hill.
A binocular view (pictured below) toward Chena
Ridge from various points showed that its crest, upon which the
white FAA structure sits, showed that a ridge closer to Fairbanks
just about, but not completely, masks the view.
Looking southwest from
radio location
Only a radio to radio test will prove the matter
out.
MONDAY - IN THE FIELD
After determining the best roof position, Mike
Willett remained there while Wendy Davis and I drove to Chena
Ridge, seeing if we could find the road which obviously led to
the top, used by the FAA to service its facility. We found the
road, but as we expected, it had a locked gate. So loading ourselves
down with a DRG115 radio, cables, a battery and 10 dB Yagi as
well as 6 dB Omni antenna, and one of two Motorola voice radios,
binoculars, and my usual cell phone, we hiked up the road, which
was an easier grade than the steep climb the day before.
We quickly were able to establish voice contact
with Mike, back on the roof, 6 miles away, and tested a Freewave
radio pair. They linked up easily, even with a six inch 'rubber
duck' antenna through light trees. This proved we would have no
trouble getting a strong link from the university to the relay
peak - provided we could get permission to use it. Wendy Davis,
who is a marathon runner, went swiftly back down the road to the
vehicle to get a cable that we needed.
Wendy Davis
with the Freewave radio and battery.
Me testing the
radio in the overgrown road we would have to put a mast tower
on.
Then Mike Willett accompanied Stephanie Pike
who had fetched the power boat from the Forest Service storage
facility with a UAF truck, as they went to the boat launch location
about 3 miles down river and set out, with Mike carrying a radio
already set up to link to the radio I carried, battery, and antennas
with connectors. And his cell phone, with which we might be able
to communicate even if the Motorolas were out of range, me calling
a Texas area code, he a Colorado one - in mid Alaska. We established
Motorola contact pretty quickly.
It took some time for Stephanie
to navigate the channels to the furthest data logger points FP4A
and FP5A, after which we tested the radios, and again got reasonable
links. Especially considering that we did not have tree-clearing
mast antennas at either end. But the results were favorable enough
for us to conclude that we also could reach those two sites -
over 11 miles from Chena Ridge, and another 6 to the University.
Proving, among other things, that the Freewave radios have excellent
performance characteristics in the field.
Then the pair visited several other data logger
sites - some getting poor links from deeper in the trees off the
river, but good links from the shore line, where the boat gets
tied up closest to the data logger's location.
FP4A data logger
site, with walkway over the spongy earth.
Site is a quarter
mile from the river.
After two hours we had learned about as much
as we could - namely that Chena Ridge was a good, probably the
best, relay site back to the University.
But Mike was able to test another method we
thought would be useful. Since Stephanie travels down the river,
summer and winter, she could carry a smaller radio, and a laptop
running the PC208W software, matching radios at the data logger
sites - none more than a quarter mile into the woods. The World
Wireless 900 Hopper Radios, or even test radios we tried out on
Sunday, much less powerful - and a whole lot cheaper - than Freewaves,
might work.
So Mike tested these at several sites and it
was clear the method could be useful for some sites that we could
not reach from the relay hill.
Our conclusions were that - each data logger
site down the Tanana River should have a crank-up, tethered, mast
that one person could carry in and erect, that would top the trees
in the vicinity of the data loggers. These cost less than $500.
They could save a large amount of labor in fetching data in winter
weather.
TUESDAY IN THE FIELD
Encouraged by the performance of the Freewaves
in the vicinity of Fairbanks, we then teamed up with Trevor White,
an undergraduate student employee of the Water and Environmental
Research Center familiar with the area, to drive the 35 miles
to the Poker Flat-Caribou Creek watershed study area. And see
what it would take to connect up two data loggers down in the
valleys in the watershed to the CR23X Data Logger weather station
on the top of Caribou Peak, from whence it would be very straightforward
to link that location with the Poker Flat Rocket Firing Facility
and its Ethernet-Internet link to the university.
We insured we had two plastic boxes with which
we could put our radios, batteries, and other paraphernalia and
secure them to the back rack on the ATVs as we bounced over the
permafrosted ground and rutted trails which I experienced on the
last trip to Caribou Peak.
We checked out the four ATV four wheelers where
they are parked at the Rocket Facility.
Luckily, the Director of the facility, was
there, and I was able to reconfirm that we could connect up our
radios to his site. I also got USGS Maps of the immediate area,
and he pinpointed the exact location of his relatively new facility
that was not on maps yet.
We then stopped at the Rocket firing 'Block
House' with its soil and grass covered roof where we could put
our base radio. From there we have an unrestricted view of Caribou
Peak, 4 miles away. It would only take a small tripod-held antenna
at that point.
We then set out to cross the highway and enter
the watershed area, which is protected from unwanted vehicular
traffic by a locked gate at the first bridge.
We first stopped at a 'confluence' location
that gives the LTER researchers problems every year. Ice begins
to build up right at a narrow stretch of creek, crossed by a low
bridge. Unless steps are taken the ice buildup creates a dam,
that forms a large ice lake that is very risky to cross with the
necessary ATVs to get into the areas beyond. At the bridge the
ice can reach 8 feet deep. The water, unable to sink through the
permafrost, flows 'over' the ice and deepens it. There are some
field structures at this location, and the grad student showed
us how high the ice can get on those buildings. The roofs are
never reached. A unique 'hot water' solution is used to prevent
the ice from daming up at the exact location. An improbable pipe
about 6 foot high, with a hand turn faucet stands close to the
bridge. The hot water is generated by a small boiler in one of
the buildings. While I could not visualize why it works, apparently
running that hot water just as the ice starts to form, prevents
the buildup at that location, leaving the route across the low
area passable.
What Kenji, the Japanese researcher had asked
if I could do on my earlier visit, was to put a digital camera,
linked wirelessly back to the University there so they could see
the junction day to day as winter closes in, from 35 miles away,
and pick the right time to apply their solution.
Confluence bridge is
at left where the ice forms.
We saw that this should be easy. We then all
visited a date logger site on Little Poker Creek sitting in the
planned burn area which was part of the watershed.
The burn area data logger
site
It didn't look good to link this site, pretty
far down in the valley, to the top of Caribou Peak. Lots of trees,
and perhaps a hill mass. Even a climb up the 30 foot tower at
this data logger site did not reveal Caribou Peak up above.
Wendy Davis and I left Mike and Trevor behind,
and we proceeded up the ever-rougher trail to the top of Caribou
Peak.
Once there I remarked to Wendy that I was not
optimistic about reaching either the Little Poker Creek or confluence
sites from the immediate vicinity of the weather and data logger
station, because it was back at least 50 meters from the 'brow'
of the hilltop. An antenna back on the data logger structure,
unless a tall tower were erected might not peep over the brow
of the hill enough to connect to radios down in the valleys.
I could not have been more wrong.
The minute Mike and I talked by Motorola radio-phone
we got a good link to their location out of sight down toward
the burn area. We ran several experiments, one in which Mike pointed
his antenna 180 degrees from Caribou Peak toward the slopes of
the burn area. The reflection of his Yagi directional signal was
good enough to get the link.
Mike on burn area tower
So, comfortable with that link, Mike's team
drove their ATV's back to the confluence data logger site, and
we tested from there. No problem.
So, to my surprise, even though the weather
station and data logger were back off Caribou Peak we got a good
link to the two critical data logger locations. And later at the
confluence location. Big pleasant surprise.
We want to put the radio as close as practicable
to the CR23X location so it can act as a 'Slave/Repeater' station,
and while repeating the signal from the data loggers down in the
valley, also be able to fetch data from the logger right close
to it, with only a 6 to 8 foot serial cable attached. Otherwise
we would have to put a complete radio with tower at the brow of
the hill, then another set at the data logger site - only 50 meters
away (unless we laid an exposed cable on top of the rocky ground
that 50 meters). With duplicate solar panels, batteries, boxes,
and all that impedimenta.
Me with crash helmet
at Caribou Peak Data Logger
The site sits back from
military crest about 100 meters
So the not-quite-line-of-sight radio links
demonstrate, once again, those intending to make a radio network
installation, must actually test every link in the field, and
not rely just on theory or calculations.
With that set of tests, we returned to Fairbanks,
pleased that it will be much simpler to get three data loggers
in the Poker Flat/Caribou Creek watershed linked up wirelessly
to the Rocket Launch facility, where we still have to work out
how best to convert the serial signals to Ethernet/TCP/IP ones.
That will probably require use of Campbell
Scientific's new NL100 device, and 3.2 version of their PC208W
software, and a tricky interface, to make it work.
COLD TEST
On Thursday, the 31st, the field site survey
was over. Mike Willett flew back to the lower 48, while I did
several exit briefings, became more familiar with the internal
data-management operations of the LTER, and made arrangements
with Stephanie Pike, who brings the data loggers modules back
from the field and transfers the data into computers on the University
network, to set up a 'Cold Test.'
The Cold Test, suggested by Dr. Terry Chapin,
PI of LTER, would be to see how our selected radios - Freewave
DRG115Ws would stand up under a normal cold Fairbanks winter.
Since this test would be aimed at seeing at
what point of cold - temperature - the radios might become erratic
in their performance, or fail. It would only be necessary to set
up a test close by the University to get an initial idea.
My plan was this:
- 1. Put a Campbell (or Pelican) type outdoor box on the
roof of the Arctic Health Center building where our 'base'
radio would be emplaced later.
- 2. In the box emplace a CR21X data logger from the LTER's
supply - a replacement spare, and 12 volt battery.
- 3. Connect up the CR21X to several temperature sensors
- at least 1 inside the box near the radio position, and 1
outside to measure open air temperature. Program the data
logger to record the temperature on a recurring basis.
- 4. Put the Freewave DRG115W I left there, inside the box,
connecting it via an SC932 DTE to DCE adapter, to the CR21X
CS/IO port by means of a standard Campbell RS232 cable. Also
connect it to the battery, and a Yagi antenna, which would not
be linking any distance to another radio.
- 5. Run a 12 volt-carrying line through a hole in the trapdoor
on the roof, into the box to the battery, and inside the building,
to a 110v outlet for a Magnum 3amp recharger. So the battery
on the roof would be constantly charged by the Magnum - which
has a good shut off circuit and 'charged' status light on
it. This way the LTER would not have to get a full permit
for a 110v circuit to the roof, since only 12volts would be
carried on the exposed line. A circuit could also be established
between the battery in the box and the data logger, with corresponding
code in the software in the data logger, to monitor the battery's
strength through the winter also.
- 6. Put a second Freewave Radio, with a 3dB 'rubber duck'
antenna, power supply, and RS232 9 pin cable inside the building
right next to the computer used by Stephanie to download,
via serial cable, the data logger modules. Configure this
radio as point to point master, and the one on the roof, as
point to point slave, and configure their call books to recognize
each other. Because of the short distance to the roof above
the second floor, the radios could be expected to communicate
with each other through the walls.
With this arrangement, once in place, it would
be a simply daily chore to connect the master radio to the PC
running PC208W software via the serial cable, and download the
temperature data, plotting it on a curve. It would take less than
5 minutes to do this each time.
Both the temperature outside, and inside the
box (where the slight heat generated by the data logger and radio
might make a difference, could be recorded, right up to the point
the radio started acting erratically, if it does.
With that plan expressed to Stephanie, who
was so experienced with the field data logger sites, I was confident
she could assemble and operate everything, without a further trip
there by us:
a. Left all the radios, connectors, adapters, all
configured properly in the LTER lab.
b. Authorized Stephanie to expend up to $500 from
our NSF Wireless funds to purchase the box, battery, and any
connectors necessary, and contract if needed for the wire run
through the roof.
c. Promised to send a Magnum battery recharger.
d. Instructed her on how to connect up the radios,
and get the system going.
So the Cold Test will be performed.
In the exit interview with Terry Chapin, PI,
I brought him up to date with the Site Survey Tests we completed,
their favorable outcome, but the critical requirement that the
LTER, and University if necessary, push to get BLM permission
to put the Relay Radio on the Chena Ridge Hill we had selected.
That I would be sending a formal letter detailing the technical
matters.
I expressed concern about what to expect in
the 'data management' of the data which would be flowing next
summer wirelessly from the data loggers - bringing it in at much
greater rate than their current manual system. That I was expected
by the NSF to help link the data to the Internet, perhaps through
their LTER Web Site. But that work - at least detailed planning
- would have to get started in the winter to that end.
I also was able to inform Terry, that an Internet
Satellite Company had expressed a willingness to install a satellite
base station, if my project purchased it, at Delta Junction -
100 miles south of Fairbanks where Dr. Chapin had a series of
field tests operating but which I could not connect to the university
network in any other way. He was delighted at the possibility.
With that, I departed Fairbanks on one of the
first overcast days, heralding the oncoming winter when no deployment
of a wireless network of radios was going to be possible before
spring.
Stephanie Pike:
a data manager's busy place of work in the Arctic Health Center
below the planned
base radio antenna site.
