Marshall bustled during the 1970s and 1980s with radar development and other projects by NCAR and NOAA. When the Atmospheric Technology Division began consolidating at the Foothills Lab in 1991, most of NCAR's Marshall staff came to FL. This winter two NCAR initiatives have brought a surge in NCAR's Marshall activity. ATD has been developing a specialized surface weather station for an Asian field project, while the Research Applications Program conducts a comparison of snow-measuring devices.
The collection of instruments at Marshall is a curious-looking one. There are nine precipitation gauges and seven remote-sensing instruments scattered around the site for studying winter weather. Each gauge is surrounded by one or two large shields, made of wood, metal, and plastic, that spread around and above the gauge like two-meter-wide flowers. The shields are meant to block wind, forcing snow to settle into the gauge rather than skirting across the top of it.
"You can tell the difference if you stand out here in a high wind," says Jeff. "If you don't have a shield, the wind creates an updraft and you don't get as much snow in the gauge." As much as 25 to 50% of a wind-driven snowfall can go unrecorded if a gauge is unshielded.
Surprisingly few places have taken the time to evaluate snow gauges and shields as RAP is now doing. Less than half of the regular National Weather Service (NWS) stations have adequate wind shielding for their precipitation gauges, according to Jeff. Precipitation records could be misleading due to the mixture of data from shielded and nonshielded gauges.
RAP is testing three different gauges and five wind-shield designs. The most imposing shield is a two-layer construction from the University of Wyoming that spans a four-meter-wide area. "It seems to work the best," says Jeff. He found that the outer shield reduces the ambient wind by 50% and the inner shield by another 25%. However, he adds, "The problem is that it's not very practical. You couldn't set this up at an airport. If it ever self-destructed, you'd have two-by-fours flying all over the runway."
The gauges themselves also vary. One of the three being tested is the current NWS standard. Others are from Norway and from a Fort Collins entrepreneur. The NWS design uses a spring scale that weighs the liquid from snowfall. In contrast, the Norwegian device uses a collection bucket, the weight of which alters a high-pitched frequency that can be converted into a measurement of liquid-equivalent precipitation amount. The bucket is filled with antifreeze and oil to melt the snow and keep it from evaporating before a reading is taken.
Another link in the snow-measurement chain is the relationship between snow crystal type and the amount of liquid water it holds. Jeff has been collecting snow-crystal photos with a contraption that he and Roy designed. Snowflakes fall through a tiny window onto a black, revolving plate that moves beneath a camera. Images are taken every four seconds and saved on videotape. Later on, the types of flakes falling at a given time will be correlated with snowfall rates and visibilities. Three different visibility sensors are being tested at Marshall. There's also an upward-pointing radar to measure snowfall velocities. "Rain usually falls at about ten meters a second [four miles an hour]," notes Jeff, "but snow falls at different speeds due to the crystal type and density structure."
Looks can deceive when it comes to snowfall. Dense, small flakes may carry as much water as larger, fluffier flakes without reducing visibility as much. Roy found last year that the nonintuitive relationship between visibility and water content may be a factor in several major airline crashes of the past few years. This winter's data at Marshall will help to clarify the relations between snow-crystal type, visibility, accumulation, and liquid water content. Date from the Doppler Next-Generation Weather Radar (NEXRAD) radar will also be factored into the equation. "We still can't determine snowfall intensity from radar very well, but we are working on a real-time calibration of the radar data with the snow-gauge measurements from Marshall and from the Chicago area," says Jeff.
RAP is finding that weather radar--even with its limitations--can provide useful data to airlines that have to make time-critical decisions in winter weather conditions. The NCAR/RAP Ground Deicing/Snowfall Display has proven very beneficial to the airlines and the municipal operations staff at O'Hare this winter. The system consists of a network of snow gauges and weather stations in the Chicago area and data from the located southwest of O'Hare. Many thousands of dollars in flight-delay costs may have been saved on a single day, 26 January, when the RAP system accurately detected a delay in snowfall arrival at O'Hare. The results from Marshall and O'Hare will help the FAA determine how snowfall nowcasts might be implemented at airports nationwide.
The Surface Sensing Group in ATD, led by Tom Horst, has spent much of its time at Marshall over the past two years testing a prototype of the third-generation PAM station. A subset of the PAM-IIIs is called Flux-PAMs, due to their inclusion of sensors to measure the vertical transport of heat, water vapor, momentum, and radiation. (Users can also add sensors for other fluxes, such as trace gases.)
To be sure, this GAME requires some strategy. Several enhancements are being developed for the GAME version of the Flux-PAM. For instance, Tom says, "they're interested in the total water budget, so we're adding a separate system for continuous measurement of soil moisture at multiple locations." Traditional soil-moisture measurement is cumbersome and labor-intensive. This device, called a time-domain reflectometer, will send electromagnetic waves through the soil along two parallel rods. By measuring the transmission time, it will discern the dielectric constant and, in turn, the total volume of water in the soil. To resolve the vertical distribution of soil moisture--which can vary greatly depending on precipitation patterns over time--the moisture readings will be taken at several levels through the topmost meter of soil.
Meanwhile, an infrared thermometer will take the soil's surface temperature. Above that, two hygrothermometers (one standard to the Flux-PAM, one added for GAME) will allow vertical temperature and humidity gradients to be measured, serving as backup for the direct measurements of heat and water-vapor fluxes.
Another revision to the Flux-PAM involves radiation. Though all of the Flux-PAMs will measure net radiation flux, GAME scientists need the four numbers that make up that flux: the short-wave and long-wave components of radiation in both upward and downward directions. (Short-wave radiation is mainly in the form of incoming solar energy, while long waves are primarily emitted upward from the ground.) For these measurements, ATD is reproducing a four-component radiation system already used with success in its Atmosphere-Surface Turbulent Exchange Research Facility (ASTER).
Finally, there are changes to the data storage and transmission systems. Removable "flash memory cards"--analogous to floppy disks, but with far more capacity--will be installed on each PAM. In real time, data will be sent from the PAMs through the Japanese meteorological satellite system. Since PAM-III uses American weather satellites for U.S. data transmission, adapting to the Japanese satellite will require a change in such variables as the length, format, and frequency of data sent. "All these things have to be juggled to come up with the optimum data transmission rate. We're working pretty intensively with the GAME scientists to specify a format that satisfies both the scientific requirements and the regulations of the Japanese Meteorological Agency."
PAMs are tough creatures, but GAME will really test their stamina. "One of the locations where they'll be deployed is Siberia, so we'll need to make sure the system can operate in the extreme cold of that environment." At latitudes that far north, the usual reliance of PAMs on solar power is also a question mark. Fortunately, the stations for GAME will be in place for at least five years, so Tom expects that power lines can be installed for them (or stations might be sited near existing lines).
The Flux-PAM prototype for GAME goes to Japan's Tsukuba University in April. Over the following three years, if all goes well, ATD will be building from 10 to 20 additional Flux-PAM stations for GAME, along with a dozen or so units for the NSF deployment pool housed at ATD.
In these days of uncertain U.S. government funding, international programs like GAME are likely to be an increasingly valued blessing for groups like ATD. Says Tom, "This project could be quite a boost for PAM-III development." --BH
"The official recording site has been moved a number of times since 1898, which will certainly affect the measurements," says Kelsch. Crosstown differences in accumulation are apparent during big storms. Another problem is that the standard measurement frequency is only once a day, typically at nightfall. "Snowfall may have settled, sublimated, or partially melted on many occasions before the early-evening measurements," says Kelsch.
March is still officially Boulder's snowiest month, but some recent autumn snows have put November at the top of Kelsch's 1986-96 records with an average of 19.0 inches. Boulder's current cooperative observing site is at the National Institute of Standards and Technology campus on Broadway. If you'd like to help out with daily observing there, or if you have a reliable record of snowfall at your home to contribute, contact Kelsch, 497-6719, firstname.lastname@example.org.