Accuracy of precipitation measurements for hydrologic modeling

Abstract

The use of precipitation data as input for conceptual hydrologic models has enhanced the need for measurements more representative of 'true' precipitation. Precipitation input to continuous watershed models is generally some form of mean basin precipitation estimate based on point measurements. Each point measurement can have large catch deficiencies due to wind, especially for solid precipitation. A brief review is made of past results from studies concerned with these deficiencies. New curves based on current studies are presented for wind-caused gage catch deficiencies for both rain and snow. The results of using gage catch correction factors to adjust precipitation input to a conceptual hydrologic model are presented.The use of precipitation measurements as a major input for conceptual hydrologic models has enhanced the need for point and areal measurements that are more representative of 'true' precipitation. Hydrologic modeling studies indicate that one of the most important factors in successful hydrologic simulation is reliable and representative precipitation data. Raw precipitation data are often converted to some form of mean basin precipitation (MBP) estimate for use in a hydrologic model. Present hydrologic models to a large degree are limited by the accuracy of the MBP estimate. Many factors influence the estimate of MBP, including density and arrangement of the network, the particular site and gage characteristics at each location within the network, methods of areal analysis utilized, basin characteristics, storm characteristics, etc. For solid precipitation the most important of all these items, however, is the gage catch deficiency due to wind [Peck, 1972]. Many studies in the past have attempted to evaluate gage catch deficiencies in terms of the causes and magnitudes of the errors [Green and Helliweli, 1972; Warnick, 1956]. In this paper a brief review is presented of some of these past results. Recent relationships for wind-caused measurement errors for both liquid and solid precipitation are presented. Finally, the results of using precipitation correction factors based on these relationships as input for a conceptual hydrologic model and a calibrated watershed are presented. REviEw OF PREVIOUS STUDIES CONCERNED WITH GAGE CATCH DEFICIENCIES Literally hundreds of articles have been published on this subject from the mid-eighteenth century to the present. Kurtyka [1953], Israelsen [1967], and Larson [1971a] have each published comprehensive literature reviews containing a total of some 1600 references in the general field of precipitation measurements. More recently, the World Meteorological Organization [1973] has published an annotated bibliography in the same subject area.Although most studies vary with one another as far as the magnitudes of gage catch deficiencies due to wind are concerned, they all reach the same general conclusions. That is,

Comparison of Snow Water Equivalent Estimated in Central Japan by High-Resolution Simulations Using Different Land-Surface Models

“…Second, the 3.3 km grid was not fine enough to resolve micro-scale convective clouds and narrow snow band (Murakami et al 2003); thus, water that should be precipitated from small cumulus over the lowlands on the Sea of Japan side of the study area would be modeled as falling in the mountainous area. Third, under-recording of precipitation by rain gauges is much higher for snow than for rain, and increases with increasing wind speed even when there is a wind shield installed (Larson and Peck 1974). Thus, precipitation recorded under windy conditions may have been erroneously low.…”

Streamflow Modelling: A Primer on Applications, Approaches and Challenges

“…Inaccuracies in precipitation data are often cited as serious impediments to successful hydrological modelling (e.g., Larson and Peck, 1974;Yang et al, 2004). In addition to typically systematic instrument error, in-situ observations of liquid and especially solid precipitation can be negatively affected by deflection of particles by the wind, though this problem can be mitigated through careful site selection (Larson and Peck, 1974). Precipitation observations in mountainous terrain can be influenced by icing or blowing snow, which can also introduce huge random errors in measurement (Lapin, 1990;Yang and Ohata, 2001).…”

A process to estimate net infiltration using a site-scale water-budget approach, Rainier Mesa, Nevada National Security Site, Nevada, 2002–05

“…Precipitation measurement error is greater for solid (snow) than for liquid precipitation (rain) (Larsen and Peck, 1974;Groisman and others, 1991;Yang and others, 1996). Corrections for precipitation type (liquid, mixed, and solid) are identified by air temperature associated with transitions in precipitation type.…”