• Ackley, S. F., , and K. Itagaki, 1974: Crystal structure of a natural freezing rain accretion. Weather, 29 , 189192.

  • American Meteorological Society, 2004: Weather and highways. Atmospheric Policy Program Policy Forum Report, Amer. Meteor. Soc., 33 pp. [Available online at http://www.ametsoc.org/atmospolicy/documents/WxHighwaysFinalReport.pdf.].

  • ASCE, 2002: Minimum Design Loads for Buildings and Other Structures. American Society of Civil Engineers, ASCE 7-02, 376 pp.

  • Baumgardner, D., , and A. Rodi, 1989: Laboratory and wind tunnel evaluation of the Rosemount icing detector. J. Atmos. Oceanic Technol., 6 , 971979.

    • Search Google Scholar
    • Export Citation
  • Bennett, I., 1959: Glaze, its meteorology and climatology, geographical distribution, and economic effects. Quartermaster Research and Engineering Center Tech. Rep. EP-105, Environmental Protection Research Division, Natick, MA, 217 pp.

  • Campos, E. F., 1999: On measurements of drop size distributions. Top. Meteor. Oceanogr., 6 , 2430.

  • Chang, J. S-T., , and J. J. Fanning, 1998: Method and apparatus for detecting a road pavement surface condition. U.S. Patent 5852243, U.S. Patent Office, 7 pp.

  • Claffey, K., , K. Jones, , and C. Ryerson, 1995: Use and calibration of Rosemount ice detectors for meteorological research. Atmos. Res., 36 , 277286.

    • Search Google Scholar
    • Export Citation
  • Cortinas, J. V., , B. C. Bernstein, , C. C. Robbins, , and J. W. Strapp, 2004: An analysis of freezing rain, freezing drizzle and ice pellets across the United States and Canada: 1976–90. Wea. Forecasting, 19 , 377390.

    • Search Google Scholar
    • Export Citation
  • Dover, J., 1999: Time-sequenced ice accretion maps for January 1999 icing event. Raytheon ITSS Presentation to National Weather Service W/OSO14, 70 pp.

  • ETI, 2005: Model SIT-6E pavement-mounted snow and ice sensor. Environmental Technology, Inc., 2 pp.

  • Federal Aviation Administration, 2003: FAA approved deicing program updates, winter 2003–2004. Appendix 4, Flight Standards Information Bulletin for Air Transportation FSAT 03-01, Order 8400.10, 56 pp.

  • International Standards Organization, 2001: Atmospheric icing of structures. ISO 12494:2001(E), 56 pp.

  • Jones, K. F., 1998: Comparison of modeled ice loads in freezing rain storms with damage information. Proc. Eighth Int. Workshop on Atmospheric Icing of Structures, Reykjavík, Iceland, RARIK Iceland State Electricity, 163–168.

  • Jones, K. F., , and A. B. Peabody, 2002: The application of a uniform radial ice thickness to structural sections. 10th Int. Workshop on Atmospheric Icing of Structures, Brno, Czech Republic, EGÚ Brno and Jihomoravská energetika, a.s., 4 pp.

  • Jones, K. F., , A. C. Ramsay, , and J. N. Lott, 2004: Icing severity in the December 2002 freezing-rain storm from ASOS data. Mon. Wea. Rev., 132 , 16301644.

    • Search Google Scholar
    • Export Citation
  • Katz, D. I., 1993: FRESNOR: A new smart pavement sensor. Transp. Res. Rec., 1387 , 147150.

  • Mulherin, N. D., 1998: Atmospheric icing and communication tower failure in the United States. Cold Reg. Sci. Technol., 27 , 91104.

  • NWS, 1995: Freezing precipitation algorithm. Automated Surface Observing System Contract 50-SANW-0050, Appendix I, Section 6.3.1.1, C-A-I-6-9.

  • NWS, 2000: Add ice accretion remark to METAR/SPECI Reports. NWS Engineering Change Proposal S01126, NWS/OM22 and NWS/OSO1x4, 45 pp.

  • NWS, 2003: WFO winter weather products specification, A4–A6, Eastern Region (NWS ER). Supplement 02-2003 to National Weather Service Instruction 10-513. 5 pp.

  • Office of the Federal Coordinator for Meteorological Services and Supporting Research, 2002: Weather information for surface transportation. NOAA/U.S. Department of Commerce, National Needs Assessment Report FCM-R18-2002, 302 pp.

  • Office of the Federal Coordinator for Meteorological Services and Supporting Research, 2005: Depth of freezing or frozen precipitation. Surface weather observations and reports. Section 12.7.2a(3), Federal Meteorological Handbook No. 1 (FMH-1), 12-14–12-15.

  • Poots, G., 1996: Ice and Snow Accretion on Structures. John Wiley and Sons, 331 pp.

  • Ramsay, A. C., 1997: Freezing rain detection and reporting by the Automated Surface Observing System (ASOS). Preprints, First Symp. on Integrated Observing Systems, Long Beach, CA, Amer. Meteor. Soc., J65–J69.

  • Ramsay, A. C., 2000: Surface ice accretion rates from the Automated Surface Observing System (ASOS): An issue for deicing holdover times. Preprints, Ninth Conf. on Aviation, Range, and Aerospace Meteorology, Orlando, FL, Amer. Meteor. Soc., 312–316.

  • Ramsay, A. C., 2002: Freezing drizzle (FZDZ) identification from the Automated Surface Observing System (ASOS): Status of the ASOS multi-sensor FZDZ algorithm. Preprints, Sixth Symp. on Integrated Observing Systems, Orlando, FL, Amer. Meteor. Soc., 241–247.

  • Ramsay, A. C., , and M. E. Laster, 1995: Status of the ASOS freezing rain sensor. Preprints, Sixth Conf. on Aviation Weather Systems, Dallas, TX, Amer. Meteor. Soc., 460–465.

  • Ramsay, A. C., , and C. C. Ryerson, 1998: Quantitative ice accretion information from the Automated Surface Observing System (ASOS). Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 502–506.

  • Ramsay, A. C., , and J. Dover, 2000: Freezing drizzle identification from the Automated Surface Observing System (ASOS): Field evaluation of a proposed multi-sensor algorithm. Preprints, Ninth Conf. on Aviation, Range, and Aerospace Meteorology, Orlando, FL, Amer. Meteor. Soc., 303–308.

  • Rasmussen, R., and Coauthors, 2006: New ground deicing hazard associated with freezing drizzle ingestion by jet engines. J. Aircr., 43 , 14481457.

    • Search Google Scholar
    • Export Citation
  • Raytheon Information Technology and Scientific Services, 1999: Ice accretion algorithm development, winter 1998–1999. Final Report to the National Weather Service ASOS Program Office, W/OSO14, 15 pp.

  • Raytheon Information Technology and Scientific Services, 2000: Compendium of ASOS icing events, 1998–2000. Report to the National Weather Service ASOS Program Office, W/OSO14, 91 pp.

  • Ronco, J., 1998: Data acquisition and availability. National Weather Service Internal Report: January, 1998, freezing rain storm, Gray, Maine, NWS WFO, 4 pp.

  • Ryerson, C., 1990: Atmospheric icing rates with elevation on northern New England mountains, U.S.A. Arct. Alp. Res., 22 , 9097.

  • Ryerson, C., , and K. Claffey, 1996: Efficacy of ice detector hoarfrost observations. Proc. Fourth Annual Mt. Washington Observatory Symp. Focus 2000: Wind, Ice, and Fog, North Conway, NH, Mount Washington Observatory, 45–55.

  • Ryerson, C., , and A. Ramsay, 1997: Quantitative glaze accretion measurements from the Automated Surface Observing System (ASOS). Preprints, First Symp. on Integrated Observing Systems, Long Beach, CA, Amer. Meteor. Soc., J71–J75.

  • Ryerson, C., , K. Claffey, , and G. Lemieux, 1994: Surface hoarfrost measurement and climatology. Proc. 51st Eastern Snow Conf., Dearborn, MI, Eastern Snow Conference, 121–130.

  • SAIC, 2001: Quick-time report of ASOS icing event at Kansas City International (KMCI), February 23, 2001. Report to National Weather Service, W/OST32, 9 pp.

  • SAIC, 2003: Case study of icing event at Denver, Colorado (KDEN), October 31–November 1, 2002. Report to National Weather Service, W/OST32, 14 pp.

  • Smithson, L. D., 2005: Surface transportation weather issues for the 21st century. Proc. 2005 Mid-Continent Transportation Research Symp., Ames, IA, Iowa State University. [Available online at http://www.ctre.iastate.edu/pubs/midcon2005/.].

  • Stein, P., 1993: Rosemount freezing rain sensor. Proc. Sixth Int. Workshop on Atmospheric Icing of Structures, Budapest, Hungary, Hungarian Electrotechnical Association, 121–124.

  • Tattelman, P., 1980: Field tests of a surface ice accretion measurement system. Air Force Geophysics Laboratory, Surveys in Geophysics Number 435, AFGL-TR-81-0007, 22 pp.

  • Tattelman, P., 1982: An objective method of measuring surface ice accretion. J. Appl. Meteor., 21 , 599612.

  • Wade, C. G., 2003: A multisensor approach to detecting drizzle on ASOS. J. Atmos. Oceanic Technol., 20 , 820832.

  • View in gallery

    Goodrich Sensor Systems (Rosemount) Model 872C3 icing sensor used in this development project and at all ASOS stations equipped with ice detectors.

  • View in gallery

    Ice rack, cylinders, and horizontal plate representative of installations used for manually monitoring ice accretion.

  • View in gallery

    Relationship between NFC and ice mass accumulated on ice-rack horizontal aluminum rods.

  • View in gallery

    Relationship between NFC and ice thickness accumulated on ice-rack horizontal plates.

  • View in gallery

    Relationship between calculated Teq and NFC.

  • View in gallery

    Linear relationship between Req and Ti.

  • View in gallery

    The 14–15 Jan 1999 icing event at Sterling, showing responses of four icing detectors to the same event. Detectors SRDC (a) 927, (b) 852, and (c) 166 were collocated. The (d) KIAD ASOS icing detector was located at the Dulles International Airport ASOS, approximately 7 km from the SRDC units.

  • View in gallery

    The 14–15 Jan 1999 icing event at Sterling, showing measured ice mass increase during the storm, and ice mass increases predicted from each of the four ice detectors. Note that the best matches are among the ice detectors that are collocated at SRDC. Correlations between the measured ice mass and the predicted masses were all greater than r = 0.98, with standard errors of less than 37.3.

  • View in gallery

    Ice detector NFC signature for three types of icing events: (a) freezing rain, (b) freezing drizzle, and (c) frost, normalized by time.

  • View in gallery

    Map of ice-storm progress on 1–2 Jan 1999 using ASOS ice-detector response and techniques presented in this paper. Isolines show ice thickness in 0.25-cm (0.1 in.) intervals (Dover 1999).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 101 101 29
PDF Downloads 77 77 28

Quantitative Ice Accretion Information from the Automated Surface Observing System

View More View Less
  • 1 Cold Regions Research and Engineering Laboratory, Engineer Research and Development Center, U.S. Army Corps of Engineers, Hanover, New Hampshire
  • 2 Hedgesville, West Virginia
© Get Permissions
Full access

Abstract

Freezing precipitation is a persistent winter weather problem that costs the United States millions of dollars annually. Costs and infrastructure disruption may be greatly reduced by ice-storm warnings issued by the National Weather Service (NWS), and by the development of climatologies that allow improved design of infrastructure elements. However, neither the NWS nor developers of climatologies have had direct measurements of ice-storm accumulations as a basis for issuing warnings and developing storm design standards. This paper describes the development of an aviation routine/special weather report (METAR/SPECI) remark that will report quantitative ice thickness at over 650 locations during ice storms using new algorithms developed for the Automated Surface Observing System (ASOS). Characteristics of the ASOS icing sensor, a field program to develop the algorithms, tests of accuracy, application of the algorithms, and sources of error are described, as is the implementation of an ice-thickness METAR/SPECI remark. The algorithms will potentially allow freezing precipitation events to be tracked with regard to ice accumulation in near–real time as they progress across the United States.

Corresponding author address: Charles C. Ryerson, Cold Regions Research and Engineering Laboratory, Engineer Research and Development, Center, U.S. Army Corps of Engineers, 72 Lyme Road, Hanover, NH 03755-1290. Email: charles.c.ryerson@erdc.usace.army.mil

Abstract

Freezing precipitation is a persistent winter weather problem that costs the United States millions of dollars annually. Costs and infrastructure disruption may be greatly reduced by ice-storm warnings issued by the National Weather Service (NWS), and by the development of climatologies that allow improved design of infrastructure elements. However, neither the NWS nor developers of climatologies have had direct measurements of ice-storm accumulations as a basis for issuing warnings and developing storm design standards. This paper describes the development of an aviation routine/special weather report (METAR/SPECI) remark that will report quantitative ice thickness at over 650 locations during ice storms using new algorithms developed for the Automated Surface Observing System (ASOS). Characteristics of the ASOS icing sensor, a field program to develop the algorithms, tests of accuracy, application of the algorithms, and sources of error are described, as is the implementation of an ice-thickness METAR/SPECI remark. The algorithms will potentially allow freezing precipitation events to be tracked with regard to ice accumulation in near–real time as they progress across the United States.

Corresponding author address: Charles C. Ryerson, Cold Regions Research and Engineering Laboratory, Engineer Research and Development, Center, U.S. Army Corps of Engineers, 72 Lyme Road, Hanover, NH 03755-1290. Email: charles.c.ryerson@erdc.usace.army.mil

1. Introduction

The Automated Surface Observing System (ASOS) has become the primary surface weather observing system in the United States. Over 880 ASOS have been commissioned; over 650 ASOS sites have received the Goodrich Sensor Systems (formerly Rosemount) 872C3 icing sensor (Fig. 1), providing the system with the ability to report freezing rain (Ramsay and Laster 1995) but with no capability to provide quantitative reports of ice accretion. The ASOS is currently programmed to report icing events only when they are associated with freezing rain. However, the ASOS icing sensor is also known to detect ice accretion from freezing drizzle, wind-driven mist that freezes on elevated objects, freezing fog, and hoarfrost (Ryerson and Claffey 1996; Ramsay 1997; SAIC 2001, 2003; Wade 2003).

A new algorithm, based on raw data from the icing sensor, has been demonstrated to provide reliable quantitative estimates of ice accretion; this new algorithm has been approved for system-wide implementation (NWS 2000). Software changes that will allow the ASOS to begin providing ice accretion reports in the aviation routine/special weather reports (METAR/SPECI) have been completed, but the date of the software release has yet to be determined by the National Weather Service (NWS). The purposes of this paper are to describe how the algorithm was created, to illustrate how ASOS icing reports may be translated to estimates of ice loading on structures, and to explain the validity and limitations of the ASOS icing reports in METAR/SPECI.

Laboratory and field tests of the technology used in the ASOS icing sensor (a magnetostrictive oscillator that responds to changes in mass on a small sensing element) (Tattelman 1980, 1982) indicated that quantitative estimates of ice accretion could be derived from sensor data. The NWS originally procured icing sensors for the ASOS in the early 1990s with the objective of identifying freezing rain, but did not actively pursue the development of a quantitative ice accretion capability. As the ASOS matured and the reliability of the icing sensors was demonstrated, the NWS undertook a formal product improvement initiative in 1995 to investigate the ability of the sensor to provide acceptable estimates of ice amounts. Field studies conducted during the winters of 1996 through 2002 as part of the product improvement initiative confirmed and refined the ability of the ASOS icing detector to derive and report quantitative ice amounts.

a. Surface-icing applications

There are many needs for quantitative ice accretion information. For example, the NWS requires indications of glaze ice accumulation for public safety and winter-storm warnings. Quantitative measures of ice accumulation have not been available on a standard surface to either human or automated weather observers. Despite this, NWS ice-storm warnings in the NWS Eastern Region, for example, are based upon quantitative accumulations of ice thickness, with thresholds of accumulated ice amounts ranging from 0.64 cm (0.25 in.) to 1.27 cm (0.5 in.) (NWS 2003). This lack of information was particularly critical during the severe January 1998 freezing-rain storm in northern New England (Ronco 1998).

Aircraft- and runway-deicing decisions are currently based upon operator judgment and not upon freezing-precipitation measurements. In freezing-precipitation conditions, anti-ice fluids are applied to aircraft after deicing is completed; these fluids provide a “holdover time” during which new ice accretion is prevented. Holdover time duration—the time between fluid application and takeoff—is a function of fluid concentration, air temperature, and precipitation type and rate. Quantitative information about precipitation rate and icing intensity could allow users to make more effective use of published holdover tables (Federal Aviation Administration 2003). In addition, as indicated by Rasmussen et al. (2006), the timely transmission of surface-icing information from the ASOS may also prevent extensive damage to jet aircraft engines.

State and municipal highway departments require indications of ice amount for deicing highways. Road weather information systems utilize roadside weather stations to report conditions to road crews. Though roadway icing is reported by embedded pavement sensors, augmentation by ASOS observations of icing conditions may be useful (Katz 1993; ETI 2005; Chang and Fanning 1998). For example, Smithson (2005) indicated that one of the problems identified by the Federal Highway Administration was the poor availability and quality of real-time precipitation rate and accumulation information. Freezing-precipitation accumulations are the most important weather impact upon most transportation sectors according to a survey conducted by the Office of the Federal Coordinator for Meteorological Services and Supporting Research (2002). Also, an American Meteorological Society (2004) policy forum reported that improved road weather observations should include precipitation type, rate, liquid equivalent, and start and end times.

Quantitative icing measurements are critical for establishing engineering-design climatologies for communication towers and power lines (ASCE 2002). Ice storms cause the collapse of communication towers (Mulherin 1998) and high-voltage transmission lines. The American Society of Civil Engineers (ASCE) is the primary source for minimum design loads for buildings and other structures (ASCE 2002). Icing design guidelines for many years were acquired principally from Bennett (1959), using icing measurements from a variety of organizations, but using no standard methods. An ASCE task committee on atmospheric icing recently developed design guidance based upon models of freezing precipitation that utilize climatological data as input (Jones and Peabody 2002; ASCE 2002). Though considerable verification of models has been accomplished using newspaper and other accounts (Jones 1998), standardized measurements, such as could be provided by more than 650 ASOS icing sensors, may provide more reliable standards and better model verification. The utility of objective and quantitative ice-accumulation estimates in a December 2002 ice storm was illustrated by Jones et al. (2004).

b. Ice-thickness measurement

One of the most serious problems about freezing rain that confronts observers is how to measure the amount accumulated. Glaze ice accretions vary significantly over short geographic distances, but also with the shape and orientation of structures on which the ice accretes (Jones and Peabody 2002), the thermal properties of those structures, and small-scale local variations in wind speed and direction. However, most important is that it is difficult to measure ice amount, even on structures as simple as tree limbs or wires. Though freezing rain occasionally creates nearly uniform ice cylinders around objects such as twigs and grass blades (Ackley and Itagaki 1974), it more often is influenced by wind, thermal conditions, and rainfall rate, allowing water to flow before freezing, creating icicles and other nonuniform shapes. The extreme complexity of ice accretion on structures is well illustrated in Poots (1996), and the International Standards Organization’s (2001) standards on atmospheric icing on structures.

The U.S. federal government has defined ice thickness as the vertical depth of ice on a horizontal surface (Office of the Federal Coordinator for Meteorological Services and Supporting Research 2005). However, news media and other sources often measure and report the maximum thickness of ice observed on any surface, including tree branches, wires, fences, and any other object that serves to dramatize the event. Every surface is exposed differently to the wind and precipitation; therefore, acquiring measurements that are representative of any other surface or that can be compared with other locations is difficult. Jones (1998) solved this problem by cutting tree limbs, weighing the ice on the limb, and computing a uniform radial ice thickness using an assumed ice density after removing the weight of the limb. Though such techniques are logical alternatives to thickness measurements when used in the research world, they are difficult to apply to the operational environment.

It should be possible for users who require quantitative ice-loading information to develop transfer functions from a standard surface to their operational interests if it could be shown that the ASOS ice detector provides representative quantitative estimates of ice accretion on “standard” surfaces. Ramsay (1997) and Ramsay and Laster (1995) demonstrate that the ASOS ice detector (Stein 1993) can reliably indicate the presence of freezing rain. Ryerson et al. (1994) and Ryerson and Claffey (1996) indicate that an early version of the ASOS ice detector is capable of quantitatively indicating hoarfrost deposition. Claffey et al. (1995) showed that an older model of Rosemount ice detector, which is of similar technology to the ASOS instrument but is intended for mountaintop work in rime icing conditions, can provide estimates of cloud liquid water content and can quantitatively indicate ice amount. These studies suggested that the technology used in the ASOS freezing-rain detector may also be able to provide quantitative indications of glaze accretions in freezing-drizzle or freezing-fog events, and with wind-driven mist (Office of the Federal Coordinator for Meteorological Services and Supporting Research 2005) particles that freeze on elevated objects. The event documented by Rasmussen et al. (2006) illustrated the ability of the sensor to provide ice accretion information in freezing-drizzle or freezing-mist conditions.

2. ASOS surface-icing project

a. Project goals

Between 1995 and 2002 the NWS ASOS Program Office, NWS support contractors [Hughes STX, Raytheon Information Technology and Scientific Services, and the Science Applications International Corporation (SAIC)], the U.S. Army Corps of Engineers Cold Regions Research and Engineering Laboratory (CRREL), the NWS Eastern Region, the National Oceanic and Atmospheric Administration National Climatic Data Center (NCDC), the Goodrich Sensor Systems (formerly Rosemount Aerospace) company, and the Mount Washington Observatory combined resources to demonstrate the capabilities of the ASOS ice detector to provide quantitative ice measurements. A goal of the project was to report ice accretion amounts from a standard surface. The ASOS icing sensor, which has minimal problems with icicles and exposure, provided the necessary standard surface for which an algorithm could be developed for representing ice accretion on other surfaces.

The Goodrich Sensor Systems Model 872C3 sensor (Fig. 1) was used in this development project. Within the ASOS, this instrument is known as a “freezing rain” sensor; however, because this development project is ultimately intended to extend the capabilities beyond freezing rain, this sensor will be referred to as the ASOS icing sensor.

The ASOS icing sensor detects ice by sensing ice mass on a 25-mm-long by 6-mm-diameter vertical cylindrical probe that vibrates longitudinally at a nominal 40 kHz when ice free. Full operating characteristics of the ice detector are described by Ramsay (1997), Ramsay and Laster (1995), and Stein (1993).

Software within the ASOS polls the icing sensor once each minute and controls sensor deicing (heating) cycles. A deicing cycle is triggered by any of three conditions: 1) if sufficient ice mass accretes that the probe frequency decreases to 39 474 Hz, then deicing will occur to prevent the probe from being overwhelmed by ice; 2) if the probe frequency decreases at a rate equal to or less than 13 Hz in 15 min, then deicing will occur in anticipation that freezing precipitation is ending because icing is so light; and 3) if “clamping” causes an anomalous frequency increase above 40 020 Hz (Ramsay and Laster 1995), then deicing will occur. Following a deicing cycle, the probe typically cools below freezing (and resumes the reporting of ice accretion) in less than 5 or 6 min; infrequently (e.g., with ambient temperature very near freezing, very light precipitation, and low wind speeds), the sensor has been known to require up to 30–45 min until the probe cools again below 0°C. Cooling rate varies with air temperature, wind speed, and precipitation rate. The probe is sensitive to any type of ice that adheres to its surface, the most typical being glaze (Ramsay 1997), rime (Claffey et al. 1995; Ryerson 1990; Baumgardner and Rodi 1989; Tattelman 1982), and hoarfrost (Ryerson and Claffey 1996; Ryerson et al. 1994).

The initial ASOS algorithm for freezing rain was developed in the late 1980s and was implemented in 1995 with the installation of icing sensors throughout the system. The ASOS relies on three sensors to define the occurrence of freezing rain: the thermometer, the precipitation identifier, and the icing sensor. The ASOS freezing-rain algorithm is written in terms of “ice thickness,” which is determined by monitoring the minute-to-minute frequency of the icing sensor and then multiplying the frequency decrease (from 40 000 Hz) by a manufacturer-specified ice-thickness factor of 0.000 152 in. (0.000 386 cm) Hz−1 (NWS 1995).1 The authors have been unable to find formal documentation or supporting empirical data for the ASOS ice-thickness factor; therefore, an additional objective of the icing project was to assess the “goodness” of this value.

The ASOS definition of ice thickness is taken to mean the depth of ice on a horizontal surface, in accordance with standards established in the Federal Meteorological Handbook No. 1 (FMH-1; Office of the Federal Coordinator for Meteorological Services and Supporting Research 2005). ASOS ice depth is used as an artifice to define the onset of an icing event at 0.005 in. (0.012 cm) of ice accretion and to define two conditions at which the sensor is to be deiced: 1) 0.08 in. (0.203 cm) of ice accretion or 2) an accretion rate of less than 0.008 in. (0.02 cm) h−1 during a running 15-min window.2 If the sensor-derived ice accretion is 0.005 in. (0.012 cm) or greater and if the ice accretion rate is 0.008 in. (0.020 cm) h−1 or greater during the preceding 15 min and if the temperature is less than 2.8°C and if the precipitation identifier reports either liquid or “undifferentiated” precipitation, the ASOS releases a report of freezing rain. The intensity of freezing rain is defined by the precipitation identifier and not by the rate of ice accretion.

The 1995–2002 ASOS surface-icing project investigated the relationship between ASOS ice-detector performance and actual ice measurements near the detector, with an objective of developing transfer functions to predict ice accretion from the icing-sensor signal. The project required an understanding of icing-sensor performance, development of an ice accretion measurement process at representative locations, analysis of paired values of sensor reports and icing measurements, and the development of reliable algorithms to transfer ASOS reports of inches of ice depth to estimates of ice accretion/loading on objects or structures. Each of these elements is described below.

b. Icing-sensor performance

The ASOS icing sensor responds to ice accretion by reporting a change in frequency. In general, sensor frequency decreases linearly with increasing ice mass. The ASOS icing algorithm relies on monitoring sensor frequency change on a minute-to-minute basis. A 1-min frequency change (decrease) is defined as the difference between the lowest frequency in the previous 15 min and the current minute’s frequency value. Continuously summing the minute-to-minute decreases of frequency over the duration of the event yields the net frequency change (NFC) in hertz. There are two anomalous situations for which the ASOS software must make allowances in the computation of NFC.

1) Anomalous frequency increases

Anomalous frequency increases are believed to be related to ice “bridging” between the base of the sensor probe and the supporting strut; periods with frequency increases are known as “clamping” events. Bridging/clamping can change the vibration characteristics of the probe and results in frequency increases, even with increasing ice mass. Short-term frequency increases are seen primarily at the end of icing events when the ice on the probe begins to melt or during wet snow events when melting ice/snow slides down to the base of the probe. Short-term anomalies are removed from the raw data by calculating frequency changes relative to the lowest frequency value in a moving 15-min window; only frequency decreases relative to that lowest frequency are included in the computation of NFC. Long-term clamping events are rare, but, when they occur, they result in the loss of reporting capability.

2) Probe temperature above freezing

During the period immediately following a deicing cycle, the probe temperature may remain above freezing, and the sensor will not be sensitive to continuing freezing precipitation. The ASOS algorithm accounts for this condition by maintaining a running value of the mean ice accretion rate over the preceding 15 min and using that mean value immediately following a deicing cycle, up to the time at which a frequency decrease is again detected, for a maximum of 15 min.

c. Icing measurements

Because direct measurement of ice accreted on trees, bridges, towers, and transmission lines is impractical for reasons described earlier, surrogate surfaces upon which to accrete ice and to measure its mass and thickness were developed and placed near ASOS ice detectors for this study. Measurement techniques used in this project were developed before the definition of international “standard” techniques (International Standards Organization 2001). Ice mass intended to be representative of accretion on trees and power lines was monitored on 32-mm-diameter, 1-m-long, thin-wall aluminum cylinders suspended horizontally on a rack (Fig. 2). Cylinders were intended to emulate high-voltage transmission lines and tree limbs and were exposed in pairs, with one serving as a backup. Racks were manually rotated during an icing event to keep the cylinders orthogonal to the wind to achieve maximum possible ice accretion because the icing sensor’s vertical probe is always oriented into the wind for maximum ice accretion. At observation time, each cylinder was removed, weighed, and returned without damaging accreted ice. Cylinders were suspended nominally 1.5 m above the ground surface and were not adjusted in height for changing snowpack depth.

Ice mass and thickness were also measured on a 20.3-cm2 by 3.2-mm-thick horizontal aluminum plate suspended also about 1.5 m above the soil surface from one end of the rack (Fig. 2). Although not thermally coupled to the ground, which is necessary to represent road surfaces (except bridges), plate ice thickness may represent ice on objects that are at or near ambient air temperature. Data from the aluminum plate suffered from problems related to the fall angle of the particles (a function of particle size and wind speed), from contamination from ice pellets and snow that do not normally adhere to the icing-sensor probe, and especially from the inability to make precise measurements of the rough ice thickness. Cylinders and plates were nominally leveled to prevent water runoff if freezing was slow. Ice mass on cylinders and plates was weighed on an electronic scale to the nearest gram. Ice thickness on the plates was measured with calipers to the nearest 0.8 mm.

Tests requiring dedicated (full time) “clinical” observations were performed primarily at the NWS Sterling, Virginia, Research and Development Center (SRDC) and at Johnstown, Pennsylvania (KJST). At SRDC, sensor data were obtained from ASOS units and from icing sensors in a sensor test bed. At Johnstown, data were obtained from the KJST ASOS and from icing sensors in a special NWS test bed for winter testing of ASOS sensors. Additional observations were made by the NWS Weather Forecast Offices (WFO) at Binghamton, New York (KBGM), and Cleveland, Ohio (KCLE); by CRREL in Lebanon, New Hampshire (KLEB); and by Mount Washington Observatory in New Hampshire (KMWO). Hourly observations were made at Sterling and Johnstown, and 2-h observations were made at other locations because the primary responsibilities of personnel did not permit more frequent observations.

Observations contributing to the dataset were made in a wide range of icing conditions and are assumed to be representative of typical icing situations occurring at airports in the United States. A problem in data collection was with wind speeds, because wind speed has a significant impact on ice accretion. However, a complete and consistent set of wind observations was not available because of the effect of ice on the cup-and-vane anemometers available at the test sites. Especially in cases with a large amount of icing, reported wind speeds are known to be reduced by the added ice mass on the anemometers. In some cases, anemometers were completely frozen and wind speeds of “zero” were recorded.

Multiple sensors were located at Sterling and Johnstown, and ice detectors occasionally were moved or replaced as failures occurred at all sites. During the four winters of observations, data were obtained during approximately 188 h of icing conditions. Five of the six observing locations yielded usable data; no significant icing occurred at KLEB. The project obtained 178 measurements of ice mass on cylinders, 126 measurements of ice mass on horizontal plates, and 133 measurements of ice thickness on horizontal plates. Combining these field measurements with contemporaneous values of sensor NFC from a mix of 20 different icing sensors, the following datasets were derived: 537 “data pairs” of cylinder ice mass and NFC, 467 data pairs of horizontal-plate ice mass and NFC, and 423 data pairs of horizontal-plate ice thickness and NFC (Table 1).

The event with the most coverage (seven sensors) occurred at Johnstown on 14–15 January 1999. The longest event (26 h), described later, occurred at Sterling on 14–15 January 1999 and was covered with three sensors.

d. Data analysis

Analysis of the data pairs is presented in the following sequence: 1) correlation between ice mass on cylinders and the NFC from the sensors, 2) derivation of equivalent uniform radial ice thickness Req, 3) comparison of ASOS-reported ice depth with measurements of ice thickness and mass on horizontal plates, and 4) transfer of ASOS-reported ice depth to Req for estimating ice accretion and ice loading on elevated surfaces.

1) Ice mass on cylinders versus NFC

Computed values of NFC for each icing sensor in each event were compared with measurements of ice mass on the aluminum cylinders. The relationship between the cumulative ice mass per unit length on the cylinder and the NFC is (Fig. 3)
i1558-8432-46-9-1423-e1
The correlation coefficient r between cumulative cylinder ice mass and NFC is 0.98, r2 is 0.96, and the standard error of estimate of ice mass is approximately 31 g m−1.

2) Equivalent uniform radial ice thickness

Observed values of ice mass can be converted to a standard representation of ice loading, the equivalent uniform radial ice thickness Req. The Req is used to characterize the mass of ice on cylindrical objects, such as conductors, wires, and power, phone and cable television lines (ASCE 2002), structural elements and guys of lattice towers, and branches and twigs of trees (Jones 1998). Our observed values of ice mass can be expressed in terms of Req by assuming that the ice is uniformly distributed in an annular volume around the cylindrical object. For each measurement of ice mass, Req is that value that gives the observed mass of ice using an assumed ice density,
i1558-8432-46-9-1423-e2
where ρi = 0.9 g cm−3 is the density of glaze ice, d is the diameter of the cylinder, and l is the length of the cylinder.
The application of the quadratic formula to Eq. (2) yields the following expression for Req in terms of the measured ice mass:
i1558-8432-46-9-1423-e3
Each of the 537 data pairs of ice mass on the 32-mm-diameter cylinders was converted to Req, using Eq. (3). The revised dataset of Req and NFC data pairs could then be further examined for utility to a wide range of users.

3) ASOS “ice depth” versus observed ice thickness, mass, and Req

Equations (1)(3) provide relationships between the NFC of the ASOS icing sensor and two specific parameters of ice accretion: the measured ice mass per unit length on a cylinder and the derived Req. However, the ASOS will report ice accretion in terms of yet another parameter—ice depth, or horizontal planar ice thickness, Ti—in compliance with the requirements expressed in the FMH-1 and in recognition of long-standing NWS procedures.

ASOS algorithms currently use the NFC of the icing sensor to calculate a value of ice depth (thickness) Ti according to a manufacturer-specified linear relationship (NWS 1995):
i1558-8432-46-9-1423-e4
Data from the horizontal metal plate used on the ice rack (Fig. 2) confirmed the expected linear relationship between ice depth and the NFC of the sensor. The relationship is very “noisy,” primarily because of the difficulty of making thickness measurements of the rough ice (Fig. 4):
i1558-8432-46-9-1423-e5
Because of the difficulties in making physical measurements of ice thickness on the plate, an alternative planar ice-thickness parameter analogous to Req is presented. This new parameter, the equivalent uniform planar ice thickness Teq, can be derived directly from ice mass measurements by assuming that the ice is uniformly distributed over the area of the plate (413 cm2) and by specifying a mean density for glaze ice (0.9 g cm−3):
i1558-8432-46-9-1423-e6
Applying this definition to the 467 horizontal-plate ice-mass data pairs, a linear regression between the calculated Teq and NFC yields the following relationship (R2 = 0.91) (Fig. 5):
i1558-8432-46-9-1423-e7
i1558-8432-46-9-1423-e8
The value of the coefficient in Eq. (7) is identical to that derived from the direct (and excessively noisy) measurements of thickness [Eq. (5)], and the equivalent in inches is within 10% of the current ASOS ice-thickness factor (0.000 152 in. Hz−1).

The relationships in Eqs. (5), (7), and (8) do not provide sufficient incentive to change the long-standing NWS icing factor in Eq. (4), given the nearly 20% variability among sensor responses (described below).

4) Transferring ASOS ice depth to structures

The ASOS METAR/SPECI reports of ice depth Ti may be used to infer ice accretion on elevated cylindrical objects, such as wires and tree limbs, in terms of Req. Figure 6 shows the relationship between Req [Eq. (3)] and Ti [Eq. (4)] for the 537 data pairs and shows a linear expression representing the Req and Ti relationship:
i1558-8432-46-9-1423-e9
This approximation provides reasonable estimates of Req for up to nearly 2 cm (0.8 in.) of ASOS-reported ice depth and somewhat overestimates Req values for higher values of ice depth (Fig. 6). Because Ti is reported by the ASOS in inches, users requiring estimates of Req in centimeters will find that the relationship between Req in centimeters and Ti in inches becomes nearly 1:1—a useful rule of thumb.

e. Case study

A significant icing event occurred throughout the mid-Atlantic states on 14–15 January 1999. The event was well covered with continuous field observations at the SRDC, and it provides an example of the performance of the ASOS icing algorithm. Three icing sensors were in operation at the SRDC during this event; the three sensors and the ice rack were mounted in a special sensor test bed at a height of approximately 2 m, all within 5 m of each other. Data were also acquired from the operational ASOS at Dulles International Airport (KIAD), 7 km to the south. Frequency-versus-time curves for the four sensors are shown in Fig. 7. The sawtooth curves represent the ice-detector-probe frequency. Decreases in frequency are caused by accretion of ice on the vibrating probe. The vertical lines represent return of the probe frequency to 40 kHz as ice melts off the probe during deicing cycles.

The continuous curves in Fig. 8 are derived from the frequency time series for each of the sensors operating during this event at Sterling and at KIAD. For each ice detector, the minute-to-minute drop in frequency during icing is summed, creating a minute-to-minute NFC (though increases in probe frequency during deicing cycles are ignored as described earlier). The variance of ice mass computed among the four ice detectors is evident, as is their overall good comparison with measured ice mass on the ice-rack rods. However, the smaller variance between measured and predicted mass of the SRDC ice detectors is probably because they were used to create Eq. (1) and are located close to the ice rack. KIAD data were not used for equation creation because of its large distance from the SRDC, which may contribute to its larger variance.

Ice-detector signature varies with icing type (Fig. 9). Freezing rain shows numerous short-period spikes that indicate that ice accumulated rapidly and the ice detector deiced frequently. Freezing drizzle shows a smaller accumulation rate, indicated by fewer deicing cycles than are caused by freezing rain. Sensor responses to freezing fog or freezing mist may appear identical to freezing drizzle. Frost shows no deicing cycles and a slow accumulation rate typical of vapor deposition. During low accumulation rates and short events, signatures of all ice types may be similar.

f. ASOS ice accretion METAR/SPECI code

With the systemwide implementation of the ASOS software containing the icing-remark algorithm, quantitative estimates of ice accretion will be routinely reported in the remarks of the METAR/SPECI code. Ice accretion values will be calculated using the manufacturer-provided ASOS icing factor [Eq. (4)], and will represent ice depth to the nearest hundredth of an inch in accordance with federal standards. The icing group (designated by the letter “I”) will be analogous to the current precipitation-accumulation “P” group. Reports will be generated for any SPECI transmission and at the time of the hourly METAR transmission for any hour in which icing is detected. Icing amounts are cumulative during the hour, with the total estimated ice depth also reported to the nearest hundredth of an inch during the hour reported in each remark.

At the intermediate synoptic times (0300, 0900, 1500, and 2100 UTC), if icing had been detected during the previous 3-h period, the total ice depth during the 3-h period would be reported. At the mandatory synoptic times (0600, 1200, 1800, and 0000 UTC), if icing had been detected during the previous 6-h period, the total ice depth during the 6-h period would be reported. If augmentation/backup should occur for freezing precipitation, and the freezing-rain sensor is either not installed or is malfunctioning, then the remark will not be encoded automatically from the ASOS. There will be no manual backup required for this remark. Detailed information on the I group is expected to be released by the NWS before the remark begins to appear in METAR/SPECI transmissions.

3. Discussion

a. Ice-report strengths and limitations

This new capability initially enhances the utility of only a qualitative detector for precipitation type. The limitations of the ASOS icing sensor to derive quantitative estimates of ice accretion that were described must be considered in light of the current total lack of reported quantitative icing information. The limitations of the ice accretion algorithm described below were thoroughly examined by operational field offices of the NWS and were deemed acceptable, resulting in a formal directive to proceed with development and implementation of the necessary software changes (NWS 2000).

1) Intersensor variability of icing estimates

Our field evaluations of the ASOS freezing-rain sensor (Ramsay 1997; Ryerson and Ramsay 1997; Ramsay and Ryerson 1998) demonstrated that the detector’s reports of ice accretion are reliable. The ice-detector probe is sensitive to ice that couples solidly to the probe surface and is generally insensitive to other contaminants such as snow and unfrozen raindrops.

However, there may be significant differences in reported ice accretion from different sensors exposed to identical icing conditions. Differences among sensors arise from the basic mechanical response of different probes to the same mass of ice. Goodrich Sensor Systems delivered all ASOS sensors in the early 1990s in full compliance with NWS specifications, which were intended only to support the qualitative identification of freezing rain. Sensors were delivered if they passed a manufacturer’s “rate test,” in which a sensor’s response to a specific ice amount and icing rate was required to be within 20% of a nominal value. A representative sample of rate tests for ASOS sensors indicated that sensor responses were uniformly distributed around ±20% of the nominal value (D. Jackson 2002, personal communication). There is no “bell curve” for these sensors, and any given sensor is equally likely to have any response within the ±20% range.

2) Reliability of icing reports

In the later stages of this development project, the NWS sponsored a series of field evaluations of this new enhanced ASOS capability. The evaluations consisted primarily of case studies of 284 icing events across the United States, representing over 1500 h of freezing precipitation, from December 1998 through March 2000 (Raytheon Information Technology and Scientific Services 2000). That study documented several significant characteristics of ASOS icing reports, which are reported here because they are not readily available to users outside the NWS. None of the conditions described below occurred frequently in the events (Table 1) studied for the derivation of ASOS ice accretion relationships [Eqs. (1), (4), (5), (7), (8), and (9)], but it is important that users are aware of the potential limitations of the ASOS ice accretion product.

Ice accretion below reporting threshold

The basic icing-sensor-related criteria for the ASOS to declare the start of an icing event (i.e., a minimum frequency decrease of 33 Hz, with a concurrent rate of decrease of at least 13 Hz in 15 min) provide a “cushion” to ensure that the system will not issue a false alarm of freezing rain. However, the conservative definition of an icing event means that, although all significant surface-icing events are detected and reported, about 6% of all minutes during icing events have accretion amounts or rates that are less than the thresholds and are therefore not reported. The icing-event definition will not impact reports of ice thickness; once an icing event has been initiated, ice accretion during these unreportable onset periods will be included in ASOS ice-thickness estimates.

End-of-icing-period FZRA reporting

The ASOS reporting algorithm for freezing rain extends METAR/SPECI reports of freezing rain (code “FZRA”) for an additional 15 min after the icing sensor no longer detects accretion. This attribute has been operating in the ASOS freezing-rain algorithm since its inception. It is designed to account for periods of “showery” precipitation and eliminates multiple SPECI transmissions for the starting and stopping of FZRA. The extension of FZRA reporting adds about 4% to the climatology of FZRA; in general, additional ice accretion is not seen during the extension period, and so this feature of the freezing-rain-reporting algorithm does not introduce inaccuracy into estimates of ice accretion amounts.

Icing-report override

ASOS algorithms for determining “present weather” may cause some real ice accretion to be ignored: if the ASOS precipitation identifier determines that snow is occurring, any decrease in frequency reported by the icing sensor will be attributed to wet snow and no ice accretion will be reported. Approximately 1% of all actual icing minutes in this dataset of 1500 h of freezing precipitation would not have been reported because of this snow override. Note that Cortinas et al. (2004) reported that from 14% (freezing drizzle; code “FZDZ”) to 24% (FZRA) of freezing precipitation occurs concurrently with snow; the ASOS reporting algorithm would prevent the reporting of freezing precipitation if the snow were detectable by the ASOS precipitation identifier.

Clamping

Frequency increases caused by ice bridging at the base of the probe are known to affect both the definition of icing events and estimates of ice accretion. Approximately 4% of all icing periods in the 1500-h dataset evidenced clamping episodes; any ice accretion that occurs during a clamping incident lasting longer than 15 min will not be reported.

Slow deicing cycle recovery

Some icing conditions may be missed by the sensor, in particular during an icing event with temperatures near 0°C. In these conditions the sensor may fail to report ice accretion for up to 30–45 min following a deicing cycle because of slow probe cooling. These periods occur infrequently, and in general they have low amounts and rates of accretion. Approximately 2% of the 1500 h of active icing in this dataset were missed because of this effect.

FZRA and FZDZ collection efficiency

The collection efficiency of the icing probe is a complex issue. It is controlled by probe collision efficiency and thermodynamic factors. The collision efficiency—how likely it is for particles actually to strike the probe—is related to the size and orientation of the probe, the wind speed, and the drop size distribution. Smaller particles in higher winds may be deflected around the probe, and the larger particles (with greater inertia) are more likely to impact the probe. In addition, those smaller particles that actually strike the probe will tend to freeze immediately on contact, whereas larger particles may have time to run off the probe before freezing (especially near 0°C) and therefore may not be represented in the frequency changes reported by the sensor. Field datasets used to develop the ASOS ice accretion algorithm represent common midlatitude stratiform precipitation regimes, in which larger precipitation drop sizes (say, >2 mm) are seldom experienced (Campos 1999).

Representativeness of observations

The icing factor determined in this project is a single empirical value derived from a wide range of wind speeds and drop sizes in more than 188 h of icing in the northeastern United States. The limited geographic sampling area is of concern, but the project had to be conducted within practical constraints of technology, manpower, budget, and time. Also, the NWS provided permission only to work at preselected locations. Also, the discrepancy between climatological expectations of concurrent snow with freezing precipitation (Cortinas et al. 2004) and the experience during the demonstration phase of the project (i.e., ∼1% coincidence of snow and freezing precipitation) is of concern but must be accepted as an observational fact. Technological limitations in the available cup-and-vane wind speed sensors (which were significantly degraded by ice accretion during the events) and the lack of particle size sensors prevented any attempt to differentiate sensor responses by wind speed or particle size.

Ice depth on ground versus elevated structures

The formation of ice on an elevated object is derived from the total flux of liquid water particles that are collected by that object. The total flux of liquid has both a vertical (precipitation) component and a horizontal (wind driven) component. It is possible, depending on wind speed and particle size distribution, for the wind-driven component of liquid flux to be significantly greater than the precipitation component. Conditions with smaller particles and higher wind speeds can cause significantly more ice on an elevated object than may be detected on the ground (or in a precipitation gauge). The vertically oriented probe on the ASOS icing sensor is more effective in detecting accumulating ice from the horizontal component of liquid flux, and so the sensor provides a more operationally useful estimate of icing on elevated objects than can be derived from a pure “precipitation” sensor. Case studies (SAIC 2001, 2003) have illustrated events in which significant icing occurred on aircraft with no reported precipitation. The complexity of ice accretion in varying wind speeds is illustrated well in Poots (1996) and Jones and Peabody (2002). Also, note that ice accretion on runways, highways, and sidewalks may be smaller or greater than that reported by the icing sensor because of thermal coupling with the soil.

Ice accretion spatial variability

It is possible that the micrometeorological conditions of the local environment could allow ice accretion at the ASOS but not on nearby surfaces. On the other hand, local ice accretion could be significantly greater than what is measured at an ASOS. Local surface ice accretion is heavily dependent upon local wind speed and direction and upon highly variable nonmeteorological factors such as topography, object orientation, temperature history, thermal mass, and local radiation sources.

b. Capabilities provided by ice reports

1) Real-time ice-thickness estimates

The primary strength of the ice accretion remark is in providing real-time ice-thickness estimates in a standard, objective, and quantified manner. In the past, this information was not available to forecasters and other users.

2) Ice types reported

ASOS ice-thickness values will be reported at any time icing is detected, whether or not they are associated with freezing rain. In the period before a new ASOS freezing-drizzle-reporting algorithm (Ramsay and Dover 2000; Ramsay 2002) is implemented or before a new precipitation identification sensor allows the direct detection of drizzle, analysts and forecasters will be able to use METAR/SPECI reports of ASOS ice accretion, with cloudy skies and no reported precipitation, to infer the existence of freezing drizzle, freezing mist, or freezing fog but will not be able to differentiate among them. Neither will the future ASOS FZDZ algorithm (implementation date unknown) be able to differentiate among the icing sources, but it will cause the ASOS to generate a report of “-FZDZ” in order to alert users to surface ice accretion. Note that “freezing mist” is not defined in the FMH-1 and therefore is unreportable by either an observer or an automated system. Ice can accrete from a freezing mist when visibility is too high to report fog and wind-driven “minute” particles, whose vertical flux may not be sufficient to be called “precipitation,” have sufficient horizontal flux to deposit detectable ice on elevated objects. An example of such an event (significant ice accretion, requiring deicing operations on aircraft, but without reportable freezing precipitation under FMH-1 rules) was contained in the SAIC (2001) report to the NWS. Ice accretion with clear skies reliably indicates frost, which may be severe enough to require aircraft-deicing operations.

3) Ice accretion on trees and wires

By using the ASOS planar ice accretion estimates, it is possible to estimate the amount of ice accreting on trees and power lines. ASOS ice-thickness reports (ice depth on a horizontal surface; in.) in the METAR/SPECI code can be interpreted as approximations of Req (cm).

4) Storm-history maps

The icing remark allows the spatial history of storm accretion amount to be mapped either in real time or after the fact. An example of the potential mapping capabilities may be seen in a snapshot of a time history of ASOS-derived surface ice thickness Ti of an icing event over the central United States on 1 January 1999 (Fig. 10) (Dover 1999).

5) Icing climatologies

For the first time, the algorithm allows objective climatologies of surface-icing severity to be created. A widely cited climatology by Bennett (1959) was created from opportune observations that often lacked consistency among information sources. Though not without fault, the new ASOS observations will allow observations that are more regionally and temporally consistent and therefore will allow more- reliable statistics.

6) Ice accretion rates

Ice accretion rates can be derived from the NFC data series (Raytheon Information Technology and Scientific Services 1999; Ramsay 2000) and could be used to determine the intensity of freezing drizzle and perhaps differentiate among icing sources (freezing drizzle, freezing fog, and wind-driven mist particles that freeze on elevated surfaces). However, the NWS does not plan to transmit quantitative ice accretion rates; nor are there plans to provide daily or monthly summaries of ice accretion.

4. Conclusions

Surface icing from freezing rain and freezing drizzle significantly impacts society, and yet before this time there have not been methods available to measure ice accumulation rate or amount systematically and objectively across the United States. The availability of a national ASOS system with icing sensors and the general verification of ASOS ice accretion estimates through this study and multiple field demonstrations now allow the NWS to indicate icing amount within the ASOS METAR and SPECI reports. Near-real-time distribution of icing estimates will allow more accurate ice-storm watches and warnings from NWS WFOs, improved information for assessing hazards to transport and utility infrastructures, tracking of ice accumulation as storms traverse the United States, and the eventual compilation of climatologies based upon ASOS-reported ice amounts. Although ice accumulation varies significantly with location because of spatial variations in meteorological and topographical conditions and the specific thermal characteristics of the accretion surface, the ASOS ice-detection system will now provide a consistent baseline of ice-amount information.

Acknowledgments

The authors thank V. Nadolski, ASOS Program Manager from 1993 to 1999, and M. Laster, M. D. Gifford, and R. Parry of the ASOS Program Office; R. Lewis of the SRDC; Dr. G. Carter, Chief, Science Services Division, NWS Eastern Region; N. Lott, S. Doty, and A. Holbrooks of NCDC; B. Childs, R. Wnek, D. Giles, J. Dover, L. Winans, D. Kaplan, and K. Weir of Raytheon ITSS and SAIC; K. Rancourt of KMWO; D. Jackson and L. Krynski of Goodrich Sensors Systems, K. Claffey of CRREL; and NWS WFOs, including J. Waldstreicher, P. Ahnert, and R. Hudgins of KBGM, R. LaPlante, C. Spicer, R. Ebert, L. Willard, and B. Mitchell of KCLE, J. Ronco of KGYX, P. Sisson of KBTV, and W. Drag of KBOX. This project was jointly funded by all of the above organizations, at CRREL by the Work Unit “Icing and Ice Adhesion Fundamentals,” and at CRREL, Raytheon ITSS, and SAIC under NWS contract. Opinions expressed in this paper are solely those of the authors and do not represent an official position or endorsement by the U.S. government.

REFERENCES

  • Ackley, S. F., , and K. Itagaki, 1974: Crystal structure of a natural freezing rain accretion. Weather, 29 , 189192.

  • American Meteorological Society, 2004: Weather and highways. Atmospheric Policy Program Policy Forum Report, Amer. Meteor. Soc., 33 pp. [Available online at http://www.ametsoc.org/atmospolicy/documents/WxHighwaysFinalReport.pdf.].

  • ASCE, 2002: Minimum Design Loads for Buildings and Other Structures. American Society of Civil Engineers, ASCE 7-02, 376 pp.

  • Baumgardner, D., , and A. Rodi, 1989: Laboratory and wind tunnel evaluation of the Rosemount icing detector. J. Atmos. Oceanic Technol., 6 , 971979.

    • Search Google Scholar
    • Export Citation
  • Bennett, I., 1959: Glaze, its meteorology and climatology, geographical distribution, and economic effects. Quartermaster Research and Engineering Center Tech. Rep. EP-105, Environmental Protection Research Division, Natick, MA, 217 pp.

  • Campos, E. F., 1999: On measurements of drop size distributions. Top. Meteor. Oceanogr., 6 , 2430.

  • Chang, J. S-T., , and J. J. Fanning, 1998: Method and apparatus for detecting a road pavement surface condition. U.S. Patent 5852243, U.S. Patent Office, 7 pp.

  • Claffey, K., , K. Jones, , and C. Ryerson, 1995: Use and calibration of Rosemount ice detectors for meteorological research. Atmos. Res., 36 , 277286.

    • Search Google Scholar
    • Export Citation
  • Cortinas, J. V., , B. C. Bernstein, , C. C. Robbins, , and J. W. Strapp, 2004: An analysis of freezing rain, freezing drizzle and ice pellets across the United States and Canada: 1976–90. Wea. Forecasting, 19 , 377390.

    • Search Google Scholar
    • Export Citation
  • Dover, J., 1999: Time-sequenced ice accretion maps for January 1999 icing event. Raytheon ITSS Presentation to National Weather Service W/OSO14, 70 pp.

  • ETI, 2005: Model SIT-6E pavement-mounted snow and ice sensor. Environmental Technology, Inc., 2 pp.

  • Federal Aviation Administration, 2003: FAA approved deicing program updates, winter 2003–2004. Appendix 4, Flight Standards Information Bulletin for Air Transportation FSAT 03-01, Order 8400.10, 56 pp.

  • International Standards Organization, 2001: Atmospheric icing of structures. ISO 12494:2001(E), 56 pp.

  • Jones, K. F., 1998: Comparison of modeled ice loads in freezing rain storms with damage information. Proc. Eighth Int. Workshop on Atmospheric Icing of Structures, Reykjavík, Iceland, RARIK Iceland State Electricity, 163–168.

  • Jones, K. F., , and A. B. Peabody, 2002: The application of a uniform radial ice thickness to structural sections. 10th Int. Workshop on Atmospheric Icing of Structures, Brno, Czech Republic, EGÚ Brno and Jihomoravská energetika, a.s., 4 pp.

  • Jones, K. F., , A. C. Ramsay, , and J. N. Lott, 2004: Icing severity in the December 2002 freezing-rain storm from ASOS data. Mon. Wea. Rev., 132 , 16301644.

    • Search Google Scholar
    • Export Citation
  • Katz, D. I., 1993: FRESNOR: A new smart pavement sensor. Transp. Res. Rec., 1387 , 147150.

  • Mulherin, N. D., 1998: Atmospheric icing and communication tower failure in the United States. Cold Reg. Sci. Technol., 27 , 91104.

  • NWS, 1995: Freezing precipitation algorithm. Automated Surface Observing System Contract 50-SANW-0050, Appendix I, Section 6.3.1.1, C-A-I-6-9.

  • NWS, 2000: Add ice accretion remark to METAR/SPECI Reports. NWS Engineering Change Proposal S01126, NWS/OM22 and NWS/OSO1x4, 45 pp.

  • NWS, 2003: WFO winter weather products specification, A4–A6, Eastern Region (NWS ER). Supplement 02-2003 to National Weather Service Instruction 10-513. 5 pp.

  • Office of the Federal Coordinator for Meteorological Services and Supporting Research, 2002: Weather information for surface transportation. NOAA/U.S. Department of Commerce, National Needs Assessment Report FCM-R18-2002, 302 pp.

  • Office of the Federal Coordinator for Meteorological Services and Supporting Research, 2005: Depth of freezing or frozen precipitation. Surface weather observations and reports. Section 12.7.2a(3), Federal Meteorological Handbook No. 1 (FMH-1), 12-14–12-15.

  • Poots, G., 1996: Ice and Snow Accretion on Structures. John Wiley and Sons, 331 pp.

  • Ramsay, A. C., 1997: Freezing rain detection and reporting by the Automated Surface Observing System (ASOS). Preprints, First Symp. on Integrated Observing Systems, Long Beach, CA, Amer. Meteor. Soc., J65–J69.

  • Ramsay, A. C., 2000: Surface ice accretion rates from the Automated Surface Observing System (ASOS): An issue for deicing holdover times. Preprints, Ninth Conf. on Aviation, Range, and Aerospace Meteorology, Orlando, FL, Amer. Meteor. Soc., 312–316.

  • Ramsay, A. C., 2002: Freezing drizzle (FZDZ) identification from the Automated Surface Observing System (ASOS): Status of the ASOS multi-sensor FZDZ algorithm. Preprints, Sixth Symp. on Integrated Observing Systems, Orlando, FL, Amer. Meteor. Soc., 241–247.

  • Ramsay, A. C., , and M. E. Laster, 1995: Status of the ASOS freezing rain sensor. Preprints, Sixth Conf. on Aviation Weather Systems, Dallas, TX, Amer. Meteor. Soc., 460–465.

  • Ramsay, A. C., , and C. C. Ryerson, 1998: Quantitative ice accretion information from the Automated Surface Observing System (ASOS). Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 502–506.

  • Ramsay, A. C., , and J. Dover, 2000: Freezing drizzle identification from the Automated Surface Observing System (ASOS): Field evaluation of a proposed multi-sensor algorithm. Preprints, Ninth Conf. on Aviation, Range, and Aerospace Meteorology, Orlando, FL, Amer. Meteor. Soc., 303–308.

  • Rasmussen, R., and Coauthors, 2006: New ground deicing hazard associated with freezing drizzle ingestion by jet engines. J. Aircr., 43 , 14481457.

    • Search Google Scholar
    • Export Citation
  • Raytheon Information Technology and Scientific Services, 1999: Ice accretion algorithm development, winter 1998–1999. Final Report to the National Weather Service ASOS Program Office, W/OSO14, 15 pp.

  • Raytheon Information Technology and Scientific Services, 2000: Compendium of ASOS icing events, 1998–2000. Report to the National Weather Service ASOS Program Office, W/OSO14, 91 pp.

  • Ronco, J., 1998: Data acquisition and availability. National Weather Service Internal Report: January, 1998, freezing rain storm, Gray, Maine, NWS WFO, 4 pp.

  • Ryerson, C., 1990: Atmospheric icing rates with elevation on northern New England mountains, U.S.A. Arct. Alp. Res., 22 , 9097.

  • Ryerson, C., , and K. Claffey, 1996: Efficacy of ice detector hoarfrost observations. Proc. Fourth Annual Mt. Washington Observatory Symp. Focus 2000: Wind, Ice, and Fog, North Conway, NH, Mount Washington Observatory, 45–55.

  • Ryerson, C., , and A. Ramsay, 1997: Quantitative glaze accretion measurements from the Automated Surface Observing System (ASOS). Preprints, First Symp. on Integrated Observing Systems, Long Beach, CA, Amer. Meteor. Soc., J71–J75.

  • Ryerson, C., , K. Claffey, , and G. Lemieux, 1994: Surface hoarfrost measurement and climatology. Proc. 51st Eastern Snow Conf., Dearborn, MI, Eastern Snow Conference, 121–130.

  • SAIC, 2001: Quick-time report of ASOS icing event at Kansas City International (KMCI), February 23, 2001. Report to National Weather Service, W/OST32, 9 pp.

  • SAIC, 2003: Case study of icing event at Denver, Colorado (KDEN), October 31–November 1, 2002. Report to National Weather Service, W/OST32, 14 pp.

  • Smithson, L. D., 2005: Surface transportation weather issues for the 21st century. Proc. 2005 Mid-Continent Transportation Research Symp., Ames, IA, Iowa State University. [Available online at http://www.ctre.iastate.edu/pubs/midcon2005/.].

  • Stein, P., 1993: Rosemount freezing rain sensor. Proc. Sixth Int. Workshop on Atmospheric Icing of Structures, Budapest, Hungary, Hungarian Electrotechnical Association, 121–124.

  • Tattelman, P., 1980: Field tests of a surface ice accretion measurement system. Air Force Geophysics Laboratory, Surveys in Geophysics Number 435, AFGL-TR-81-0007, 22 pp.

  • Tattelman, P., 1982: An objective method of measuring surface ice accretion. J. Appl. Meteor., 21 , 599612.

  • Wade, C. G., 2003: A multisensor approach to detecting drizzle on ASOS. J. Atmos. Oceanic Technol., 20 , 820832.

Fig. 1.
Fig. 1.

Goodrich Sensor Systems (Rosemount) Model 872C3 icing sensor used in this development project and at all ASOS stations equipped with ice detectors.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 2.
Fig. 2.

Ice rack, cylinders, and horizontal plate representative of installations used for manually monitoring ice accretion.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 3.
Fig. 3.

Relationship between NFC and ice mass accumulated on ice-rack horizontal aluminum rods.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 4.
Fig. 4.

Relationship between NFC and ice thickness accumulated on ice-rack horizontal plates.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 5.
Fig. 5.

Relationship between calculated Teq and NFC.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 6.
Fig. 6.

Linear relationship between Req and Ti.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 7.
Fig. 7.

The 14–15 Jan 1999 icing event at Sterling, showing responses of four icing detectors to the same event. Detectors SRDC (a) 927, (b) 852, and (c) 166 were collocated. The (d) KIAD ASOS icing detector was located at the Dulles International Airport ASOS, approximately 7 km from the SRDC units.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 8.
Fig. 8.

The 14–15 Jan 1999 icing event at Sterling, showing measured ice mass increase during the storm, and ice mass increases predicted from each of the four ice detectors. Note that the best matches are among the ice detectors that are collocated at SRDC. Correlations between the measured ice mass and the predicted masses were all greater than r = 0.98, with standard errors of less than 37.3.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 9.
Fig. 9.

Ice detector NFC signature for three types of icing events: (a) freezing rain, (b) freezing drizzle, and (c) frost, normalized by time.

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Fig. 10.
Fig. 10.

Map of ice-storm progress on 1–2 Jan 1999 using ASOS ice-detector response and techniques presented in this paper. Isolines show ice thickness in 0.25-cm (0.1 in.) intervals (Dover 1999).

Citation: Journal of Applied Meteorology and Climatology 46, 9; 10.1175/JAM2535.1

Table 1.

Ice mass, thickness, and NFC measurements during study.

Table 1.
1

U.S. customary units (in.) are used because they apply to NWS reporting procedures.

2

The running 15-min window was defined by algorithm developers in the late 1980s and was intended to “smooth out” system reaction to short-term sensor frequency changes; the 15-min smoothing prevents multiple transmissions of SPECI reports in showery precipitation.

Save