Winter 2002-2003
Alaska Radar Field Project

Field Project Overview

Data collection is an essential component of NCAR's Juneau turbulence project. In the absence of empirical data that allow the correlation of anemometer and radar measurements with quantifiable levels of turbulence, all the software algorithms and meteorological sensors designed and installed by project engineers and scientists would be useless. Data provides the foundation to determine whether or not a given set of wind conditions are likely to be associated with hazardous levels of turbulence. The collection of this data is the purpose of our field projects.

The scenario is simple. While our anemometers and wind profilers, a type of weather radar, collect wind data, a specially-instrumented aircraft accurately monitors atmospheric turbulence in regions known to be hazardous. Subsequently, statistical analysis of correlations between sensor data and aircraft data allow the spectrum of wind conditions likely to generate extreme turbulence to be characterized. When the turbulence warning system is deployed, software algorithms automatically process the wind sensor data and issue a warning when conditions likely to be associated with turbulence are detected.

A robust turbulence warning system requires that the underlying data represent a wide variety of conditions, including an entire spectrum of wind and turbulence conditions. Data collection can only occur when all wind sensors are operational and the aircraft is functional. The need for an extensive data set is the reason that so many trips to Juneau have been necessary. During the 2002-2003 winter field program, two aircraft were used to collect turbulence data: a KingAir leased from the University of Wyoming and a 737 provided by Alaska Airlines.


Doppler on Wheels

Although not a component of the operational turbulence warning system, the Doppler on Wheels (DOW) mobile radar was incorporated into the 2002-2003 winter field program. One of the objectives of the FY03 NCAR Juneau field program was to determine the reliability of the DOW as a surrogate for aircraft measurements of turbulence. In theory, scanning Doppler radar measurements should embody information about turbulence. Turbulence measurements acquired via ground-based radar would be considerably less expensive than in situ aircraft measurements. In addition, the DOW can safely operate during conditions that are considered too hazardous for the research aircraft to fly.

The DOW was specifically designed to observe rapidly migrating small-scale severe weather phenomena, such as tornadoes, thunderstorms and hurricanes. It is something of a celebrity and is often featured in weather-related television programs involving storm and tornado chasing. The truck and radar are a little beat up because the DOW has sustained several direct hits by tornadoes. If the DOW can chase tornadoes and thunderstorms, it can also chase turbulence, provided that one has some idea of where to look.

The DOW is an "X-Band" radar, which means that it operates at a frequency that cannot directly detect the movement of clear air. A clear-air radar would require a dish too large to house on the back of a truck. X-band radars rely upon the presence of wind-embedded particles to reflect and scatter radar signals. The motion of these particles, typically rain or snow, but sometimes insects or particulate debris, accurately represent the motion of the wind. In the presence of precipitation, DOW data quality is generally excellent. On the other hand, in the absence of rain or snow, data quality is so poor as to be essentially useless. Fortunately, Juneau winters offer abundant rain and snow. Unfortunately, virtually all DOW operations occur in inclement weather.

A small "operator's hut" is mounted just behind the truck cab. This confined shelter houses all of the computers and hardware necessary to control the radar transmitter, antenna movement, and data collection and monitoring. It can also accommodate two operators, and an additional two people can be seated in the truck cab. For most of the FY03 field program, the DOW operators were Joseph Jou and Michael Knepp. We spent many rainy and snowy hours in the DOW together.

Most radars are affected by "clutter," which refers to signal reflections from non-meteorological targets that contaminate the wind-related returns. Typically, these unwanted echoes are caused by the movement of tree branches, birds and aircraft, sea waves, etc. The DOW has a unique type of unwanted signal that I call "Bubba clutter." People are naturally curious when they see the DOW, which some believe resembles a UFO mounted on a pedestal. Not surprisingly, this curiosity was amplified considerably in a small, isolated, totally-dead-in-the-winter town like Juneau, and people routinely approached the truck to find out what was going on. Sometimes little was happening weather- or turbulence-wise and the distraction was welcome. During other periods, however, the distraction was extremely inconvenient. We used to joke about ways to keep people at bay, such as wearing biohazard suits, or posting a radiation symbol or a sign that said "Donations Welcome" on the side of the truck. Of course, we never did any of these, but we enjoyed thinking about it.

Late one night, the curiosity of the locals reached a new peak. We were deployed in the parking lot of the Wetlands Wildlife Refuge which is next to the main highway in Juneau. In addition to the multi-colored safety and operational lights that are routinely used at night, the rotating radar dish was brightly illuminated with floodlights. Apparently, this caught the attention of countless drivers and we were soon the subject of local talk radio. For an hour or so, it seemed that everyone in Juneau was at least curious, if not deeply concerned, about the "thing" at Wetlands. A few days later, the Juneau Empire published a short piece on the DOW and its role in the NCAR field program, and our fifteen minutes of fame was brought to a conclusion.


Challenges and Frustration

Any scientist or engineer with field experience is well aware that events rarely go strictly as planned. Going into the FY03 Juneau field program, I had visions of the careful and deliberate collection of large volumes of high-quality radar and turbulence data. In reality, I spent much of the field program feeling like the little Dutch boy with his finger in the dike. The DOW operations were plagued by one mechanical problem after another. The relatively harsh Alaska winter may have pushed the DOW to its operational limit, or it simply may have been time for a thorough maintenance overhaul. Whatever the reason, the continued participation of the DOW in the NCAR Juneau field program was threatened on a number of occasions.

Fortunately, there are two DOW radars - the DOW2 and the DOW3. The original DOW fell apart years ago. The DOW2 arrived in Juneau, via the Alaska Marine Highway Ferry system, in late October 2002. Within days the azimuth antenna control motor failed. Against all odds, we had the repair completed within four days. The previous time this motor failed, the DOW was out of service for more than a month. For the next few weeks, the DOW2 was affected by numerous minor problems such as a leaky roof, nitrogen leaks in the wave guide, loose cable connections, and a broken speedometer, yet remained basically functional. In mid-November, however, the radar transmitter failed.

The radar transmitter is essentially a high-voltage black box that must be repaired by the manufacturer. It could not be fixed or replaced in Juneau, the entire unit would have to be shipped to Texas. For a short time, it seemed that the DOW experiments for the FY03 field program would end. Then we decided to ship the DOW3 from NCAR to Juneau. There was one huge caveat, if we could not get the truck from Boulder, Colorado to Bellingham, Washington (north of Seattle) by Friday afternoon, we would have to wait an additional week for the next ferry to Juneau.

The transmitter failure occurred on a Wednesday night and by early Thursday morning, we had an NCAR radar technician driving the truck toward Seattle. This journey was, in part, an act of faith. We knew that a solo driver would be unlikely to reach the Washington ferry dock on time (1400 miles), so we just got him headed in the general direction with the promise that we would figure something out. We scrambled and found out that we could get the DOW operators from Juneau to Boise, Idaho, via Alaska Airlines, about the time that the DOW would be passing through Idaho. Fortunately, Alaska Airlines was an early sponsor of our turbulence project and were extremely helpful.

The DOW operators, Joseph and Michael, arrived in Boise on schedule and drove the truck to Washington. During this time, I was in Boulder doing my best to resolve as many of the logistical problems as possible. Friday morning, I received a phone call from Joseph assuring me that they would make the Bellingham ferry with a few hours to spare. He also asked me if I would use the internet to find a Denny's restaurant in their vicinity. Juneau does not have a Denny's, and after a month without Denny's, Joseph could hardly think about anything other than his favorite restaurant. If they couldn't find one in Washington, it would be his last chance until the field program was suspended for Christmas break. Fortunately, I found a Denny's and gave Joseph the exit number. He promised to call me when they arrived at the ferry dock.

Late that afternoon, I received a call from the cell phone in the DOW. Expecting to hear that they had arrived and would soon be on their way to Juneau, I was instead told that the truck was billowing smoke and gushing oil. My first thought was that this was their way of getting even for some of the practical jokes that I played on them in Alaska. Unfortunately, this was not the case. The DOW truck engine was billowing smoke and gushing oil. After driving all those miles, they could see the Pacific Ocean, but could not get to the ferry dock. We later confirmed that they were literally less than five miles from the ferry dock.

It was "panic and scramble" time again. I had them call a list of Bellingham towing services that I got from the internet while I called the Alaska Marine Highway information desk to find out if a vehicle could be towed onto the ferry. To my surprise, towed vehicles were allowed, but when I called Michael and Joseph back they were having no luck finding a company with a large enough truck to tow the DOW. At that point, a minor miracle occurred. A state patrol officer stopped to see what the problem was and, after listening to their explanation, called for the appropriate sized tow truck and even got permission for them to tow the DOW directly onto the boat so that the tow truck would not have to wait around for the loading to begin. The next time Joseph called, they were on the ferry.

We had two days to arrange a tow truck to get the DOW off the boat in Juneau. It literally took the largest tow truck in town because the ramp from the ferry to the dock is very steep. But we got it there. Now we had two DOWs in Juneau. One had a functional radar, but was not driveable. The other had a functional engine, but a useless radar. Unfortunately, the DOW radar parts are not interchangeable. We would have to work with one or the other.

Although the DOW was "semi-mobile" with the use of a tow truck, there were many problems. Radar deployments were not limited to a nine-to-five timeframe and and the tow truck was often unavilable on short notice. Also, the truck and radar could not be situated as accurately when it was manuevered by towing. The greatest challenge, however, was leveling the truck. It is absolutely essential that the truck bed be precisely level. If this were not the case, we could not know where the radar beam was located at a given time. Leveling of the radar platform is accomplished with a system of hydraulics that is powered by the truck engine. With the truck engine disabled, it was necessary to manually level the truck bed with jacks that were obtained from the local Rent-A-Center. This process could take more than thirty minutes, and the results had to be frequently monitored and adjusted when the DOW was deployed on non-paved surfaces. Given the conditions in which the DOW was typically operational, this was usually a wet and cold chore.

During Christmas break, we had the engine in the DOW3 replaced, so when we arrived back in Juneau in early January, we had a fully operational truck and radar. For the first time in more than a month-and-a-half, all the major DOW components were functional. This situation would soon change, however. About one week after our return, we began to notice a jerking sensation during rotation of the radar dish. We soon discovered that an electrical problem in the antenna control mechanism was causing the dish to lose its signal within specific azimuth ranges, which triggered the stow mechanism in the antenna controller. The "jerk" we were feeling was the dish abruptly attempting to stow and then returning to its prescribed elevation angle within a fraction of a second. Although an NCAR radar technician was dispatched from Boulder to fix the problem, we later discovered that the erratic behavior of the dish had been going on for some time. Although the loss of dish elevation control was not extreme enough to be felt, it was significant enough to cause the beam to track incorrectly. A simple antenna tracking verification algorithm would have prevented this problem, but none was installed in the DOW. Because this problem was not detected sooner, nearly one-third of DOW data was useless.


Rewards and Accomplishments

Fortunately, the Winter 2002-2003 NCAR Juneau field program was not all panic and frustration. Immediately upon my arrival for the January phase, I was fortunate enough to hitch a helicopter ride up to Sheep Mountain during an inspection of our anemometer station and survival shelter. On the way up, we flew over Douglas Island and the Gastineau Channel. The ride down surpassed even the ride up, we headed up to the Juneau Ice Field and descended directly over the Mendenhall Glacier. The return ride included a spectacular view of the Mendenhall Towers that I will never forget.

As I mentioned above, the DOW radar is useless in the absence of precipitation. So days without rain or snow were free for exploring the fringes of the wilderness that isolates Juneau from the outside world. We saw whales spouting and diving, and Bald Eagles were everywhere. Away from town, Juneau is surprisingly beautiful and peaceful in winter. The Mendenhall Glacier was covered with pristine white snow and, at times, there was not a soul in sight. Although this might seem inconceivable for someone that has only visited the glacier during the tourist season, the last cruise ship departs Juneau shortly after Labor Day and the town can seem almost deserted.

We took many hikes through forests and along beaches. We also took countless pictures. We visited all of the major tourist sites and many places that will never be known to daytrippers that arrive on cruise ships. One day Joseph and I literally drove to the end of the road, a place about 30 miles north of Juneau where travel stops unless one has a plane, boat or dog sled. We ate at our favorite restaurant, a Thai place called "Chan's." We listened to veteran pilots tell stories about flying in Alaska. Overall, life was pretty good.

Above all else, we got to know the DOW. We learned to love it, and we learned to hate it. When all systems were functional, we were operational within five minutes of arriving at a deployment site. We discovered all the tricks that the storm and tornado chasers have developed to minimize deploymentf time. We powered up the radar transmitter before driving to a site, so it would be warm when we arrived. Often, one operator would have the truck leveled before the other finished deploying the GPS receivers. We could quickly determine the truck heading from GPS signals, radar solar calibration, an analysis of a reflectivity scan, or a combination of these techniques. We collected more than 100 hours of Doppler radar data on nearly 80 DVDs, much of which was coincident with aircraft turbulence measurements. We became proficient at correlating the images we saw on the DOW radar screens with the wind, rain and snow that we felt outside the truck, as well as our understanding of the meteorology of Southeast Alaska. In short, we learned a lot and we accomplished even more.

DOW reflectivity fields were used to assign data confidence levels. Extremely low reflectivity was indicative of an absence of precipitation necessary to provide a strong return and reliable winds data. In contrast, extremely high reflectivity was associated with radar returns from solid, non-atmospheric targets, such as trees, rocks, etc. DOW radial velocity and spectral width fields were used to estimate various scales of turbulence. DOW-based turbulence estimates were correlated with aircraft measurements of turbulence.

The data collected was later used to demonstrate that the detection of atmospheric turbulence with a DOW-type radar is feasible. Unfortunately, the DOW could never replace aircraft because it doesn't work in clear air.

---

Photo Gallery