Over the three year period of the project, the number of variables being measured was expanded. Consistency was maintained from year to year, however, in terms of station location, types of sensors used at each site, and in the heights of sensors above the ground.
This section presents details on the meteorological sensors and their installations. The measurements themselves, numerical processing by the datalogger, and the subsequent adjustment and quality control of data are discussed in subsequent sections.
Air temperature measurements were all made with the same type of sensor, a Fenwal Electronics UUT51J1 thermistor incorporated into a Campbell Scientific, Inc. (CSI) model 207 probe, and housed in a naturally ventilated 12 Plate Gill Radiation Shield. (IMAGE: sensor in use) The resolution of this sensor, over a measurement range of -20° C to +20° C, with a 12 bit conversion, is 0.035° C (Tanner, 1990). The overall probe accuracy, in a worst case additive error situation, is specified by CSI as ± 0.4° C, which is probably of the same magnitude, or less than, the error associated with radiation shield limitations (Tanner, 1990). The sensor height was consistent from year to year at 180 to 190 cm above the ground.
The 1992 air temperature values have been adjusted slightly, based on a four day post-field season intercomparison of the sensors used at Delta and Echo. Using an average difference in the comparison period values of 0.16° C, Delta values were adjusted up 0.08° C, and Echo values down 0.08° C.
Water vapor measurements were of relative humidity (RH), which was measured directly, and vapor pressure, which was calculated from the RH and temperature measurements (1992 only). Relative humidity was measured by a CSI model 207 probe, comprised of a Phys-Chemical Research Corp. model PCRC-11 RH sensor, which detects vapor adsorption onto a sulfonated polystyrene surface. The probe was housed, with the air temperature sensor, in a naturally ventilated 12 Plate Gill Radiation Shield. (IMAGE: sensor in use) Because the model PCRC-11 sensor loses stability through time, the RH chip was either purchased or replaced within one month of field use, in all instruments and years except 1991. The 1991 Delta sensor was the same as that used in 1990, so was one year old. The manufacturer's stated accuracy of the sensor is +/- 1 percent, although operational accuracy has proven to be more like +/- 5 percent (Tanner, 1990). A limited number of comparisons at Taconite Inlet, between relative humidity measured by the RH sensor and handheld psychrometers, revealed acceptably small differences. The sensor height on the weather stations was consistent from year to year at 180 to 190 cm above the ground.
The 1992 RH values have been adjusted slightly, based on a four-day post-field season intercomparison of the sensors used at Delta and Echo. Based on an average difference in the values of 1.2 percent, Delta values were adjusted down 0.6 percent, and Echo values up 0.6 percent.
At the Delta station, a R.M. Young Co. model 05103 propeller-vane anemometer was used to measure wind speed and direction. (IMAGE: sensor in use) The accuracy of the wind speed sensor was within 0.3 ms^-1, with a starting threshold of 1.0 ms^-1; the accuracy of the wind direction sensor was +/- 5° , with a threshold of 1.3 ms^-1 or less, depending on displacement angle. When the sensors were installed, the height of a horizontal axis through each instrument was 240-250 cm.
Atmospheric pressure was measured with the same instrument in the same location each year. The instrument used was a CSI SBP270 Barometric Pressure Sensor, based on a Setra model 270 variable capacitance barometer. The stated accuracy was +/- 0.2 mb. A pressure correction factor (1.2 mb) was added to the offset in the micrologger program to output measurements as sea level pressure, in millibars.
Incoming solar radiation (K-down) was measured throughout the study period as the total short-wave flux on a horizontal surface. This flux is typically termed global radiation, integrating both direct and diffuse components. All measurements were made with Eppley Laboratory, Inc. Black and White Pyranometers, model 8-48, which are thermopile-type detectors. The domes on all pyranometers were Schott WG7 glass, with a manufacturers stated transparency range of 0.28 to 2.8 µ m. Eppley states the instrument has a linearity of +/- 1 percent from 0 to 1400 Wm^-2, and the following cosine response: +/- 2 percent from 0-70° zenith angle, and +/- 5 percent from 70-80° zenith angle. The instruments at all installations were kept level, at between 1.05 and 1.5 m above the snow-free ground surface. (IMAGE: sensor in use)
Reflected solar radiation (K-up) was measured in 1992, also with an Eppley model 8-48 pyranometer. Each instrument was mounted in an inverted position beneath the K-down radiometer, at one end of a horizontal support. The other end of the support was attached to a vertical shaft. (IMAGE: sensor in use) The radiometers were situated 3 to 4 m from the tripod.
Calibration of the short-wave pyranometers was done by Eppley, with respect to the World Radiation Reference (+/- 2 percent for model 8-48). The specific instruments used in 1990 and 1991 to measure K-down had most recently been recalibrated in May 1982, because prior to the 1990 field season, these instruments only had an estimated six months of cumulative use following the 1982 calibration. The same pyranometer was used at the Delta site in 1990 and 1991. In April 1992, two additional 8-48 pyranometers were recalibrated by Eppley Labs for use during the 1992 season. These two radiometers were used to conduct a clear-sky comparison with the two instruments used to measure K-down in 1990 and 1991. Measurements were made on the roof of Morrill Sciences building at the University of Massachusetts. Measurements both before and after solar noon indicated that each older instrument was within 0.5 percent of one of those recently recalibrated, closer than the recalibrated pyranometers were to one another. Therefore, each pair of "matched" instruments was used together during the 1992 field season (i.e. for K-down and K-up). The instruments recalibrated in April were done again after the 1992 field season, and not found to have changed significantly. The significance of this calibration discussion is that solar radiation measurements over the three years were both as accurate, and as precise, as possible using 8-48 pyranometers.
Through the 1992 field season, the biologically effective component of global solar ultraviolet radiation (UV-B) was measured continuously. The sensor was a model UVB-1 Pyranometer, made by Yankee Environmental Systems, Inc. (YES). The UVB-1 has a spectral response of 0.28 to 0.33 µ m, and a cosine response of +/- 5 percent for solar zenith angles less than 60 degrees. The instrument was installed 2.0 m above the snow-free ground surface, the detection surface was kept level and the dome was periodically cleaned.
Calibration of the UVB-1 Pyranometer was done by YES in April, and repeated in September, 1992. YES performs absolute calibrations several ways, including comparison with a calibrated, NIST traceable, reference detector.
Thermal radiation was measured through the 1992 field season only. Thermal irradiance from the atmosphere, or incoming long-wave (L-down), was measured directly, while thermal exitance, or outgoing long-wave (L-up), was computed from remote measurements of surface temperature.
Incoming thermal irradiance (L-down) was determined using an Eppley Laboratory, Inc. model PIR pyrgeometer. The PIR measures 4-50 µ m wavelength radiation transmitted through a coated silicon hemisphere. ManufacturerÕs stated linearity is +/- 1 percent, and for a diffuse radiation source there is an insignificant cosine response. Field measurements were made of the pyrgeometer thermopile signal (mV output), rather than through the thermistor-battery-resistance circuit incorporated into the PIR.
Field measurements of pyrgeometer case thermistor resistance were also made, using a DC half bridge (YSI model 44031 precision thermistor). These resistances were used to determine the pyrgeometer case temperature, allowing a correction to be entered into the L-down computation accounting for changes in ambient temperature.
The pyrgeometers at both sites were mounted at the end of vertical rods at 1.6 m, adjacent to the tripod. (IMAGE: sensor in use) The pyrgeometers were not ventilated, due to logistical considerations.
Calibration of the pyrgeometers was done both before and after the 1992 field season. One instrument was calibrated at the Eppley Laboratory (in April), and all other calibrations were conducted at the National Atmospheric Radiation Centre (NARC) of Environment Canada. The use of calibration factors is discussed further in the quality control section.
Thermal exitance from the surface (L-up) at each station in 1992 was computed from remote surface temperature measurement by an infrared temperature transducer, or staring radiometer. The instruments used were Everest Industries, Inc. model 4000A, mounted at 1.7 m, on or near the weather station tripod. (IMAGE: sensor in use) The field of view is 15° , resulting in a temperature measurement from 1.7 m of approximately a 900 cm^2 area. The instruments use filters to achieve a spectral band pass of 8 to 14 µ m wavelength. The manufacturers stated accuracy is +/- 0.5 deg C, with a repeatability of +/- 0.1 deg C.
Surface temperature values measured by the staring radiometers were uniformly adjusted, based on calibration at 0 deg C. The adjustment, or offset, was determined by plotting snow surface temperatures during a period of warm temperatures and uniform snow cover, when the snow was melting. The interval used was a 38-hour period from the afternoon of June 20 to early on June 22, when snow surface temperature was very consistent. Measured values indicated an offset of 2.70 deg C (+/- 0.07 S.D.) (38 hours at 60 measurements per hour yields n = 2280; deviation calculated from n = 38). This field calibration was supported by a laboratory experiment using an ice-water bath, which confirmed the sense and relative magnitude of the offset, over a shorter averaging period.
The 1991 measurement of net all-wave radiation was done with a Micromet Instruments miniature net radiometer, also referred to as a Fritschen type. This is a temperature compensated device that measures the difference between incoming and outgoing radiation. The manufacturers stated spectral response is from 0.3 to 60 µ m. The instrument had new domes installed, and was calibrated by the manufacturer, in April of 1991. The instrument was mounted approximately 1 m horizontally from the weather station tripod, at a height above the snow-free surface of 1.15 m. Radiometer domes were only irregularly cleared of all moisture following precipitation. No adjustment has been made to net radiation data to compensate for precipitation wetness effects and/or riming.
All precipitation measurements were made at the Delta site with an All Weather Rain Gauge made by Lake Region Rehab. Ind., Inc. This instrument is essentially a one-half scale version of the National Weather Service Standard Rain and Snow Gauge, Specification No. D040-SP001. The opening diameter of the gage was 10 cm, and the calibrated internal measuring tube had a resolution of 0.2 mm. The gage was situated on the tundra, with the opening 35 cm above the ground level.