E
Automated Surface Observing System
Impact on the Climate Record
The ASOS hygrothermometer (McKee et al., 1996b) is cooler than the conventional HO-83 hygro–thermometer for both maximum and minimum temperatures and also has a smaller diurnal temperature range (McKee and Doesken, 1997; McKee et al., 1996b; Schrumpf and McKee, 1996). The maximum temperature differences are larger in magnitude (compared to minimums) and vary more with varying weather conditions. Individual test sites showed wide variation in ASOS-conventional differences, possibly due to differences in instrument siting and surroundings, as well as variable changes in the solar heating effects; this local effect can vary from day to night, and the effect of instrument location change can sometimes be as large as that resulting from the change in instrument. These local effects introduce a nonlinear relationship between the ASOS and pre-ASOS data. Large ASOS-conventional differences in dew point temperature can occur from station to station, but without systematic bias (McKee et al., 1996b). The cool temperature bias of ASOS means that relative humidity estimates are slightly higher than before, with seasonal averages being one to three percent higher (McKee et al., 1996b).
The ASOS Heated Tipping Bucket (HTB) gauge consistently undermeasured precipitation compared to the standard universal weighing gauge, during heavy rain events (McKee et al., 1996b) and snow events (McKee et al., 1995). This difference showed a non-linear seasonal pattern in the central United States, with ASOS measuring significantly less precipitation during winter and summer when compared to spring and autumn (McKee et al., 1995). The HTB gauge also reported too many days with 0.01 inch resulting from dew condensation, not precipitation. ASOS precipitation undercatch ranged from two to 10 percent compared to traditional manual observation. Further, the HTB evaporated or sublimated precipitation falling below 15°F, recording almost no cold weather precipitation. The introduction into service, beginning in 1996, of a modified heated tipping bucket gauge for ASOS resulted in an improved comparison between the ASOS and conventional measurements (McKee and Doesken, 1997). However, according to the CSD, the nearly 10 years of undercatch reported from the HTB gauge is still in the extant climate record. The phased introduction of this new ASOS gauge will complicate future precipitation comparison studies and any adjustments that may be made to the data for normal computation. Further, ASOS is not equipped to measure snowfall and snow depth amounts (NWS, 1992a).
Conventional NWS wind measurements use a three-cup design with a continuous output to drive a dial indicator or a strip chart recorder; ASOS uses a light chopper rather than a voltage generator resulting in a lower starting threshold and an accurate one-second average sample speed. The conventional wind vane reports in 10-degree steps or a resolution of ± five degrees, while the ASOS wind vane reports to the nearest whole degree (Lockhart, 1995). ASOS also introduces a significant change in the way wind speed is measured. All applications of maximum wind speed which relate to warnings have been based in the past on “instantaneous” values equivalent to an averaging time of 2 seconds, whereas ASOS uses
a five-second (soon to be changed to three-second) average peak gust. ASOS provides a two-minute average, continuously updated each minute, for the hourly observation (Lockhart, 1995, 1996a, 1996b). Possible sources for differences in wind direction (Lockhart, 1996b) are that measurements may not be taken at exactly the same time, the instruments are not co-located which would affect the character of the wind flow, and the wind direction is determined differently. ASOS provides unweighted (objective) averages (scalar or unit vector) from one-second samples taken for two minutes, whereas the conventional observation is the (subjective) average direction and speed inferred by an observer watching a dial for one minute. Analysis of five-second wind averaging indicates ASOS peak winds are lower than the previous subjective measurements (Lockhart, 1996b; McKee et al., 1996a). Differences in the hourly wind speed observation show a nonlinear wind speed-dependent bias (Lockhart, 1996b). A comparison of the wind direction distributions at two sites indicated that there was no significant difference between the ASOS and conventional hourly observations (Lockhart, 1996b).
The ASOS cloud height indicator (CHI) is a laser ceilometer that differs from the standard NWS ceilometer in the way it processes returns for low cloud base and total obscuration. Both ceiling height and cloud coverage (up to 12,000 feet only) are determined by time averaging over a 30-minute period the conditions directly overhead. In manual observations, the observer subjectively evaluates the ceilometer trace at a single point in time to determine ceiling height, and the cloud coverage is determined by visual examination of the cloud conditions over the entire sky then subjectively forming a spatial average (Cornick and McKee, 1993). ASOS ceiling reports were highly correlated to conventional ceiling reports most of the time (92.7 percent), but the high level of equality drops during periods of active weather (Cornick and McKee, 1993).
ASOS is not equipped to measure sunshine duration (NWS, 1992a). Conventional pressure observations are based on an aneroid altimeter indicator or a precision aneroid barometer with observations made at hourly and special observation times (NWS, 1992a, 1994a). The ASOS barometers consist of redundant digital pressure transducers utilizing capacitive sensors, which compute and update the pressure report once every minute from readings obtained every 10 seconds (NWS, 1992a).
Manual observation of weather phenomena, including obstructions to vision, has been based on personal interpretation of the human senses (NWS, 1994a) for almost all of history (Cornick and McKee, 1993), with intensity being based on visibility criteria. These phenomena include (a) rain, snow, fog, haze, and freezing precipitation; and (b) tornadoes, funnel clouds, water spouts, thunderstorms, hail, ice crystals, snow pellets, snow grains, ice pellets, drizzle, blowing obstructions (snow, sand, dust, spray), and smoke. The automated observation of these elements required a fundamental change in observational technique and perspective. The ASOS Precipitation Identification (PI) sensor can discriminate between the occurrence of rain and snow (and identify intensity) from an algorithm based on sensor response (Cornick and McKee, 1993). Fog is reported if visibility drops below seven statute miles and dew point depression is 4°F or less. If the dew point depression is greater than four degrees and no present weather is indicated, then haze is reported. ASOS cannot report the weather phenomena from group (b) above (NWS, 1992a). In a study of 13 sites, ASOS and human observers reported approximately the same number of total minutes of freezing rain, however the coincidence rate (ASOS and human reporting freezing rain at the same time) was about 66 percent (Ramsay, 1997).