BRUCE R. ROBINSON AND JOHN S. HOOVER
BELFORT INSTRUMENT CO.
ABSTRACT
Accurate measurement of all forms of precipitation for hydrological modeling has been the subject of numerous studies over the last few decades. Many papers have described various types of gauge and wind deflector configurations and corrective models that improve the accuracy of precipitation measurement and correct for wind induced error and wetting losses and trace precipitation errors. The consensus of all of these papers is that wind related collection errors are the most significant factor in under measurement of total precipitation. Modern computer modeling of wind flow around precipitation collection devices has led to some breakthrough designs in wind deflection devices that are easy to install, cost effective and will reduce inconsistencies in collection efficiency among various gauge geometries.
1. 0 BACKGROUND
Wind induced errors in precipitation measurement have been known for over 100 years. As early as 1881 variation in catchment as a function of height above ground were well documented (Simmons 1881). He reported as much as an 80% error in given wind conditions as a function of height. Today we know that he was reporting a decrease in collection efficiency due to increased wind speed at the collection orifice as wind speed increases logarithmically with height due to decreasing surface friction effects on the air stream as it rises above ground level. In 1974 Larson and Peck reported wind induced under catchment as much as 1% for each mile per hour of wind at the height of the collection orifice.
According to the WMO/CIMO (1985) the methodology for modifying wind flow around a precipitation gauge that most closely approximates a “true” measurement is a cluster of bushes surrounding the gauge and trimmed to the same height as the gauge orifice. As this is impractical to install and maintain in most gauge locations a secondary reference methodology was established as a double wooden slat fence surrounding the gauge to facilitate accurate intercomparison of gauges, called a Double Fence Intercomparison Reference (DFIR). Although discrepancies between the DFIR methodology and the “bush” methodology existed they were modest (92 to 95% of actual snowfall) and systematic, therefore a correction factor could be established. Today the DFIR method is most universally accepted methodology for intercomparison of precipitation gauges.
As a practical matter even the wood constructed DFIR method to modify wind flow around precipitation gauges was found to require regular maintenance and has limited applicability as it requires a fairly large area, (12 meter diameter), surrounding each gauge for effective implementation. The fixed wooden slats of the DFIR are also prone to clogging with wet snow when certain conditions exist, occasionally introducing an unknown variable to its effectiveness.
Wood DFIR construction
In 1937 A more rugged and practical wind deflection methodology was devised that was self cleaning and provided some wind deflection away from the precipitation gauge, a set of equally spaced tapered metal slats suspended from a circular metal ring (designated as an Alter Shield) and positioned from 2 to 4 feet from the center of the precipitation gauge. The metal slats were allowed to swing radially in the wind to reduce the ability of snow or freezing precipitation to stick between the wind deflecting slats. Numerous studies of this methodology indicate that it substantially improves precipitation collection efficiency but falls short of the performance of gauges using the DFIR wind deflection methodology. In a 1987 Canadian test during the month of January using vintage model 3000 Belfort Universal Weighing gauges the Alter shielded Gauge showed a 22% deficiency in precipitation collection vs the DFIR shielded gauge. Although a substantial improvement over the unshielded gauge that showed a 44% collection deficiency the Alter shield shortcomings are well documented and a correction coefficient has been difficult to determine as the geometry of the shield changes randomly with wind speed increases.
Alter Shield (Note the non uniform movement of the tapered wind deflectors)
In Canada the Nipher shield has demonstrated and ability to increase the collection efficiency of precipitation gauges. It is a solid shield that resembles the bell of a horn with the wide end upright and placed over the precipitation gauge. It has demonstrated an ability to achieve collection efficiencies approaching 90% of gauges using a DFIR at wind speeds up to 9 miles per hour and 60% at 18 miles per hour. However since it is a solid material, it must be maintained regularly to prevent snow from bridging over the collection orifice or heated at considerable energy cost to reduce bridging. It is still widely used in Canada.
A look at mean annual accumulated winter precipitation (1987 to 1991) of greater the 5 mm when compared to an adjusted (corrected to true precipitation) DFIR shielded measurements showed the following:
Note: the Belfort Gauge used in this comparison is the vintage Belfort Model 3000. The standard gauge used in the intercomparison is the standard manual AES precipitation gauge
General Conclusions from the above Canadian study include:
- Automatic Gauges commonly used during that time period, the Belfort Model 3000 and the Fisher Porter without wind shielding collected between 58 and 45 % respectively of the corrected DFIR shielded gauge during snow events and slightly above 70% of total precipitation from all forms of precipitation.
- The use of Alter shields on the automated Belfort Model 3000 gauge and the Fisher Porter gauge improved snow collection by 10-16% and all types of precipitation collection by 6-10%
- The Nipher Shield used on the Belfort Model 3000 gauge and the AES manual gauge improved the collection efficiency to 80-90% of the DFIR shielded AES manual gauge.
The new standard used by the US Climate reference network is the Small Double Fence Intercomparison Reference (SDFIR). It is about 8 meters in outside diameter and achieves collection efficiencies in high wind snow events that are within 90% of a large (12 meter outside diameter) DFIR and is more practical to use at many sites. When combined with an Alter shield internal to the wooden inner shield it has demonstrated an ability to improve overall collection efficiency to a level close to the stand alone standard DFIR shields.
In 2008 Belfort Instrument Company started testing a new metal double alter shield that would be easier to install, self cleaning, quiet and as effective as the US Climate Reference Network standard SDFIR. The new Belfort Alter Shield (patent pending) consists of closely spaced rectangular metal wind deflectors that are limited in travel by rubber grommets and springs as shown in figure #3. Limiting the travel assures uniform wind deflection even in high wind precipitation conditions. Spring loading and noise suppression grommets help reduce wind deflector noise to a minimum while still permitting adequate travel to reduce possible snow accumulation.
Belfort Alter Shield (Wood SDFIR in background)
2.0 USING WIND SHIELDS TO REDUCE DIFFERENCES IN CATCHMENT DUE TO GAUGE GEOMETRY
Two major considerations in precipitation gauge design are the inlet diameter and the exterior gauge geometry.
The inlet diameter used most commonly as a standard endorsed by the WMO is 6.28 inches (200 square cm inlet area) while the most commonly used inlet diameter in the United States is the 8 inch (324 square cm) opening. One could assume that the larger the opening the larger the potential for accurate catchment in non-uniformly dispersed precipitation. Also it can be shown that the effective catchment area of a larger diameter inlet presents a disproportionately large collection area to blowing precipitation while being less prone to bridging of blowing precipitation. Why not make all weighing precipitation gauge inlet orifices larger to reduce these catchment errors? Manufacturing costs rise dramatically as the volume of precipitation that must be measured per inch of equivalent precipitation goes up by the square of the increase in inlet radius.
The gauge geometry dramatically affects the catchment efficiency of unshielded precipitation gauges. As show below typical gauge geometries vary considerably.
Field tests have shown that in wind flow deflected by the sides of the above gauges will direct air flow across the inlet orifice at different velocities and with a varying amount of vertical wind component leading to differences in catchment during blowing precipitation events.
Figure 4 below shows a comparison of a two gauges with geometries 1 and 2 inside conventional double Alter shield wind deflectors. Note: When inside a double Alter shield, gauge Geometry 1 (DA-Ref1)is collecting blowing precipitation at a better rate than gauge Geometry 2 (DA-Bel1)when wind exceeds 5 m/s in blowing snow conditions. Note that Geometry 1 (DA-Ref-1) gauge has a 6.28 inch inlet diameter and Geometry 2 (DA-Bel1) has an 8 inch inlet diameter and that the smaller inlet diameter is collecting more precipitation. Note that the yellow line indicating the difference in collection rates only increases if the wind exceeds 5 m/s during the snow event, i.e. the Alter Shield effectively minimizes geometry differences below that wind speed.
Comparison of Gauge Geometry 1 to Gauge Geometry 2 inside a Double Alter Shield
Figure #4 Comparison of Gauge Geometry 1 to Gauge Geometry 2 inside a Double Alter Shield Figure 5 below shows a comparison of two gauges with Geometries 1 (DFIR-REF4)and 2 (DFIR-BEL3) inside Small DFIR wind shields and it becomes clear that the different geometries and inlet diameters make little difference in catchment efficiency in high wind snow events when effectively shielded.
Comparison of two Gauge Geometries inside SDFIR wind shields
Figure 6 below shows a comparison of two gauges with Geometries 1 (DBS-REF3)and 2 (DBS-BEL2)inside a Belfort Double Alter shield indicating that the Belfort Double Alter shield also minimizes effect of different gauge geometry and inlet diameter during windy snow event.
Comparison of different Gauge Geometries inside Belfort Double Alter Shields
Based on these preliminary test results and subsequent field studies it is apparent that the use of either the SDFIR or the Belfort Double Alter shield significantly reduces the differences in snow catchment efficiency between gauges having different geometries and inlet diameters at wind velocities over 5 m/s.
3.0 COMPARING MODERN WIND DEFLECTORS IN BLOWING SNOW CONDITIONS
As Shown in the 1987 Canadian Study collection efficiency of weighing type precipitation gauges varies significantly during attempts to measure solid precipitation. Our tests of three types of wind deflectors during a windy snow event indicate significant improvements can be made to collection efficiency using modern SDFIR and Belfort Alter Shield designs. Figure #7 below compares at all gauges monitored during the 12/24/09 windy snow event and indicates the following:
- SDFIR shielded gauges and Belfort Alter Shielded Gauges Show an average improvement in collection efficiency over Alter Shielded gauges of 30%
- The variability in collection efficiency between different gauge geometries is significantly improved with either SDFIR or Belfort Alter Shields (note tight grouping of 4 upper sets of SDFIR and Belfort Alter Shield data vs 3 lower sets of Alter Shield data)
CONCLUSION: Previous tests of unshielded gauges clearly indicate inaccuracies, (% short of corrected DFIR enclosed gauge), in the collection of precipitation of 42 to 55% during snow conditions and 30% of total accumulation for all forms of precipitation. These studies also indicate that adding a conventional Alter Shield improves these results by 10 to 16 % in snow conditions and 6 to 10% for all forms of precipitation. Based on the above comparison testing at the NOAA Climate Reference Site at Marshall Colorado (As presented at the January 2010 American Meteorological Society Meeting) the effectiveness of newer designs such as the SDFIR and Double Belfort Alter Shields in minimizing collection efficiency differences between differing precipitation gauge geometries and collection orifice diameters has been demonstrated. Furthermore these tests clearly demonstrate that modern wind shield designs can substantially improve overall collection efficiency of automatic weighing precipitation gauges when compared to gauges using older Alter Shield designs when wind speeds exceed 5 m/s during the precipitation event.
Further comparison testing of Gauge Geometry #3 with improved Double Belfort Alter Shields is underway and will be reported at the end of the 2010-2011 test period at the NOAA/CRN Marshall CO.Test Site.
Graph comparing snow collection of SDFIR to Double Belfort Shield (DBS)and Double Alter Shield (DA)with wind plotted on right axis. Note:” Ref “one is Gauge geometry 1 “Bel” is Gauge geometry 2
BIBLIOGRAPHY
- Weiss, L.L. and W. T. Wilson, “Precipitation Gage Shields” , IASH Pub., 1957 43(1), p 462-484
- Goodison, B.F. and Louie, P.Y.T. and Yang, D, 1998, “ WMO Solid Precipitation Measurement Intercomparison, Final Report, Instruments and Observing Methods, Report No 67.WMO/TD-No. 872
- Hansen, S and Davies, M.A., “Windshields for Precipitation Gauges and Improved Measurement Techniques for Snowfall”, 2002, USDA Forest Service , TechTips
- Rasmussen, R., Baker, B, Kochendorfer, Meyers, J. et. Al. The NOAA/FAA/NCAR Winter Precipitation Test Bed: How Well Are We Measuring Snow? 8-14-2010, National Center for Atmospheric Research, NOAA, Environment Canada.
