United States
Environmental Protection
Agency
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S3-90/064  Sept. 1990
Project Summary
 Progress  Report:   Comparison of
 Precipitation  Measurements  by
 Nipher-Shielded  and Standard
 Belfort Recording  Rain Gages at
 NADP/NTN  Sites
Carol L. Simmons and David S. Bigelow
   Persistent concern by the scientific
community that wet deposition was be-
ing incorrectly calculated due to under-
estimation of snow fall prompted EPA to
install devices on  the field sampling
gagesto improve catch efficiency. Some
evidence indicated that snow fall was
being underestimated due, in part, to
windy conditions. A shield was installed
on the regular collection gage to improve
catch during snow and windy conditions.
The data collected during this study was
compared to data collected at the same
sites using the  normal system.
   This  Project Summary was devel-
oped by EPA's Atmospheric Research
and Exposure Assessment Laboratory,
Research Triangle  Park,  NC, to an-
nounce key findings of the research
project that is fully documented in a
separate  report of the same title (see
Project Report ordering information at
back).

Introduction
   A widely recognized source of error in
precipitation estimates is incomplete collec-
tion of snow by precipitation gages, par-
ticularly under windy conditions. The Belfort
Universal  recording precipitation gage, the
type used in most weekly monitoring net-
works in the United States, has been shown
to underestimate snowfall by as much as
50%,  even when equipped with an Alter
shield to deflect the wind. In recent years, a
precipitation gage known as the Canadian
MSC Nipher-Shielded snow gage has  re-
ceived considerable attention as an im-
proved gage for monitoring snowfall. This
gage has been shown to capture 90-100% of
"ground-true" snowfall under a wide range
of environmental conditions. The superior
performance of this gage  has led to its
designation as the official  snow gage in
Canada. Canadian scientists have adapted
the Nipher shield for use on standard Belfort
and Fisher and Porter recording precipitation
gages. In tests outside of Toronto, Canada,
the Nipher-modified Belfort gage collected
92-93% of the snowfall captured by the
MSC Nipher- shielded snow gage.
    In late 1987, the US-EPA installed
Nipher-shielded Belfort recording precipi-
tation gages at nine NADP/NTN sites to
allow comparison of precipitation measure-
ments from the Nipher-shielded gages with
those from the standard Belfort recording
precipitation gages already in use at the
sites. A primary objective of the study was to
assess the relative performance of the two
gages under standard NADP/NTN operat-
ing protocols. A though the Nipher shield
was designed specifically to improve snow
capture, a second objective was to com-
pare measurements bythepairedgagesfor
various types of precipitation. This informa-
tion is important for networks that utilize a
single recording precipitation gage during
sampling intervals that contain both  rain
and snow. If the existing gages are modified
for winter use by installing the Niphershield
at the onset of winter, there  is a high prob-
ability that some rain will fall afterthe Nipher
shield is in  place. This progress report
summarizes preliminary results from ap-
proximately the first year of the comparison
study.

-------
   Discussion of Results and
   Conclusions
       Weekly precipitation values from  col-
   located Nipher-shielded and standard Belfort
   recording rain  gages were analyzed for
   significant differences in volume. The effects
   of site, precipitation magnitude and precipi-
   tation type on the differences were evalu-
   ated .  In addition, the daily amounts from
   two  sites were analyzed  and correlated
   against concurrent wind speed and tem-
   perature measurements in an attempt to
   elucidate the nature of the variability  be-
  tween the two types of gages.
      Study sites were selected to represent
  a variety of snow collection conditions. The
  Nipher-shielded rain gages are  located
  within two to 15  m of the standard  rain
  gages, and are operated and maintained by
  the site personnel in the same manner as
  are the  standard NADP/NTN rain gages.
  Maintenance includes charging the gages
  with antifreeze at the onset of winter. At the
  beginning of the study, the rain gages were
  calibrated.
     Weekly precipitation totals were calcu-
  lated  by summing the daily  amounts. Daily
  records of precipitation type and amount
  were  used to classify the weekly precipita-
  tion as to type (snow,  rain, mixed or un-
  known). The weekly  precipitation was
  classified as "snow" if > 67 percent of the
  total precipitation during the week occurred
  as snow, or "rain" if > 67 percent occurred as
  rain. Other combinations were classified  as
  "mixed",  unless > 33 percent of the total
  amount was "unknown", in which case the
  precipitation for the entire week was clas-
 sified  as "unknown". Only those weeks for
 which one or both gages  indicated that
 precipitation  occurred,   for   which valid
 measurements were available from both
 gages, and for which the precipitation type
 was known were used  in the analyses  of
 weekly data. One site was excluded entirely
 because of insufficient data.
     At all sites the mean difference be-
 tween the gages was positive (i.e., the Nipher
 gage recorded more precipitation than the
 standard gage), but the bias was significant
 (P<0.05) at only five of the eight sites. The
 magnitude of  the total difference  ranged
 from 0.04 in. to 2.88 in. (water equivalent).
 On a percentage basis, the Nipher-modified
 gage measured up to 17% more precipita-
 tion (in total) than the standard gage. There
 was noclear relationship between thegages,
 nor did the presence or absence of an Alter
shield surrounding the standard gage ap-
pear to play a critical role in determining
   whether a significant difference between
   gages was observed.
       An objective of the current analysis
   was to evaluate, on a site-by-site basis, the
   effect of the precipitation type on the differ-
   ence between the gages. This analysis was
   constrained by small sample sizes for some
   precipitation types at some sites; however,
   mean differences between the gages were
   found to be highly significant (paired t-test,
   P<0.01) for snow at two sites, and for rain at
   three sites. The largest percentage differ-
   ence between the gages for snow occurred
   where the Nipher gage recorded 37% more
  snow (in total) than the standard gage. This
  percentage difference represents an abso-
  lute  difference of only  1.46 in.  (water
  equivalent). The largest percentage differ-
  ence for rain was 17%. The lack of signifi-
  cance for some site-type combinations may
  be attributable, in part,  to small sample
  sizes.
      To circumvent the problem of small
  sample sizes, the effects of precipitation
  type were also evaluated using the combined
  data from all sites. The difference between
  the gages was found to be highly significant
  for each of the three precipitation categories
  analyzed (rain, snow, and mixed); however,
  the mean differences for each of the three
  types were not significantly different from
  one another. This result is surprising, given
  the fact that previous studies  have shown
  that error in precipitation gage measure-
  ments  is greater for  solid than for  liquid
  precipitation.  The Nipher-shielded gages
  may have overestimated rainfall by capturing
 droplets that had splashed off of the shield.
 Although somefield observers reportedthat
 rain did  appear to be  splashing off of the
 shield into the gage, without "ground true"
 information it cannot be determined whether
 this phenomenon results in an overestimate
 of the rainfall.
     The examination of both weekly data
 and daily data is important to the interpre-
 tation  of the differences (or lack of differ-
 ences) between the gages. The daily data
 permit fine--resolution in theclassification of
 precipitation type, while the weekly data are
 used by NADP/NTN to calculate deposition
 and weighted-mean concentrations. Daily
 data are more likely than weekly data to be
 affected by the carry-over of snow from one
 day to the next due to the tendency for wet
 snow to stick in the gage orifice.
    The  lack of significant differences be-
tween the gages at some sites  is puzzling,
given the  fact that previous studies have
shown  that the Nipher modification en-
hances snow capture. The age of some of
   the standard rain gages in the current study
   and the conditions under which NADP sites
   are  operated may contribute to greater
   "noise" in the current study than  has been
   the case in previous studies, making it more
   difficult to detect small differences between
   the gages. Larger sample sizes should help
   address this problem . It is also possible that
   the environmental conditions at the study
   sites are sufficiently different from those in
   previous  study areas to account for the
   differences in the results.
       Uncertainties in  the  measurements
   made by the Belfort recording  rain gages
   used m this study and throughout the net-
  work may be due to a number of factors,
  with  the relative  importance of the factors
  varying fro m site to site. First, at very windy
  sites,  the  rain gage  is subject  to "wind
  shake". Wind-induced vibrations cause the
  precipitation pen to oscillate, thereby in-
  creasing the uncertainty in the  interpreta-
  tion of the chart. The Nipher-shield appears
  to increase wind shake under some condi-
  tions, causing even greater uncertainty in
  the interpretation of the Nipher gage chart
  Second, although all gages used in the
  study were calibrated atthe beginning of the
  study, the gages tend to lose their calibra-
  tion over time and most NADP site opera-
  tors  do  not  have the  necessary skill to
  calibrate the gages. Older gages may lose
  their  calibration  more rapidly than newer
  gages. Although  calibration checks were
  made during the period of the study, the
  analysis of potential impacts of  lack of
  calibration  (if any) on the  differences be-
  tween the gages was beyond the scope of
  this preliminary  analysis.  Third, because
 the cumbersome Nipher-shield must be
  removed in order to access the catch bucket
 operators may be more likely to operate the
 Nipher-shieided gage on the back traverse
 of the gage where at sites where the collec-
 tor is installed on  a platform that provides
 little room for maneuvering, as is often the
 case at sites with large amounts of snow
 Finally, because the gages  are maintained
 weekly, rather than  daily, there  is  limited
 opportunity to detect and rectify problems
 such as "capping over" of the Nipher gage
 that can  lead to inaccuracies in precipita-
 tion measurement.
    The preliminary results of this study do
 not provide convincing evidence that the
 adoption of the Nipher-shielded rain gage
 for snow  measurements would  enhance
 precipitation capture on a network-wide
 basis; the effects of the shield appear to be
 highly  site-specific. The  site-to-site differ-
ences are not clearly attributable  to differ-
ences in amounts of snow and wind.
                                                                               .S. GOVERNMENT PRINTING OFFICE:  1990/748-012/20097

-------

-------
  Carol L Simmons and David S. Bigelow are with the Natural Resources Ecology
  Laboratory, Colorado State University, Fort Collins, CO 80523
  S.M. Bromberg is the EPA Project Officer (see below).
  The complete report, entitled "Progress Report: Comparison of Precipitation Measure-
    ments by Nipher-Shielded and Standard Belfort Recording Rain Gages at NADP/
    NTN Sites," (Order No. PB90-261 538AS; Cost: $15.00 cost subject to change)
    will be available only from:
         National Technical Information Service
         5285 Port Royal Road
         Springfield, VA 22161
         Telephone: 703-487-4650
  The EPA Project Officer can be contacted at:
         Atmospheric Research  and Exposure Assessment Laboratory
         U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
 United States                             Center for Environmental
 Environmental Protection                   Research Information
                                         Cincinnati, OH 45268
Official Business
Penalty for Private Use $300


EPA/600/S3-90/064

-------