Measurements of light rain, drizzle and heavy fog

Abstract

Measurements of light rain, drizzle and settling out (or collision collection at the surfaces) of heavy fog can be important to estimate daily, monthly and annually averaged precipitation amounts. For an example, in northern climates (Stewart et al. 2004), there are no accepted standards for the short-term measurements (on the minute time scale) due to low precipitation rates. This chapter explores the potential of various instruments to measure light precipitation and settling rate under heavy fog conditions. The World Meteorological Organization (WMO) Guide to Instruments and Methods of Observation (WMO 1983) suggested that precipitation rate should be measured from 0.02 up to 2000 mm h1 and time average should be 1 min. In the guide, required uncertainties were 0.1 mm h1 between 0.2 and 2 mm h1 and 5% above 2 mm h1 (also in WMO 1983). There are several manual instruments that collect precipitation amounts based on various techniques but most of them do not satisfy the criteria stated by WMO (Sevruk and Hamon 1984; Goodison et al. 1998). Nystuen et al. (1996) provided an extensive work on quality of automatic precipitation measurements. Automatic rain gauges usually provide both accumulated precipitation amount (PA) and precipitation rate (PR). The main types of rain gauge systems include (1) tipping bucket systems, (2) weighing systems and (3) optical systems. In addition to these systems, disdrometers and radar-based systems have also been used for precipitation measurements. The tipping bucket and weighing systems (Humphrey et al. 1997; Nystuen 1999; Ciach 2003) are affected by the flow irregularities occurring within the catchments basin and flow chambers and also by time response occurring during the tipping process. Rain gauges in general are assumed to underestimate rain due to wind and turbulent effects at the edges of the rain gauges (Yang et al. 1998). Nystuen et al. (1996) stated changes in droplet size spectra could not explain the scatter observed in optical rain gauges. This suggests that rain measurements face large uncertainties that need to be explored. Tipping buckets sample rain amount differently compared to other instruments. They usually tip when approximately 0.2 mm accumulation occurs. The rainfall rate accuracy of the tipping buckets cannot be better than 12 mm h1 over a minimum interval (Nystuen et al. 1996; Humphrey et al. 1997). This means that it will not measure precipitation rates until this amount is reached and therefore cannot be used for measurements of light precipitation (<0.5 mm h 1). Weighing rain gauges work by weighing the accumulated water amount for a specific time period. They are not useful for light precipitation types e.g., drizzle and heavy fog conditions. Nystuen et al. (1996) suggested that their expected accuracy in precipitation rate can be about 1.8 mm h 1 over a 10-s sampling time period but, in reality, this can not be achieved because measurements are not possible during the draining of accumulated water. Optical gauges have also been often used for measuring precipitation rate and precipitation accumulation. The most common ones are (1) ScTI optical rain gauges (Nystuen et al. 1996) and (2) VAISALA FD12P (Sheppard and Joe 2000; Gultepe and Milbrandt 2007a,b). These instruments are based on scintillation in an optical beam produced by rain/snow drops falling between a light source and a receiver. The change in light intensity due to a drop is related to drop size, fall velocity, optical geometry and the light source. The studies of Wang et al. (1978) and Wang and Clifford (1975) suggested that variations in light intensity are related to rainfall amount and this concept is used to calculate PR and PA values (Nystuen et al. 1996). The typical PR values from optical probes can include an uncertainty between 0.2 and 0.4 mm h1. The FD12P, however, cannot be used for snow accumulation and melted equivalent of snow total amount if they are not calibrated against a weighing gauge at the field (personal communication, VAISALA Inc., 2007). The disdrometers have also been used commonly for PR and PA calculations that are based on droplet size distributions (Illingworth and Stevens 1986; Kruger and Krajewski 2002; Tokay et al. 2003). The most common ones used are Joss and Waldvogel (JW) disdrometer (1969), POSS (Sheppard and Joe 1994), OTT Parsivel (Loffler-Mang and Joss 2000) and two-dimensional video disdrometers, 2DVD (Kruger and Krajewski 2002). The OTT Parsivel measurements can be displayed over 32 size bins and stores drop counts for a given time step. The size range is between 0.062 mm and 24.5 mm in diameter. Within high sound/electrical noise environments, this threshold amount for disdrometers increase significantly, usually to more than 0.5 mm. Nystuen et al. (1996) compared the precipitation rates from various instruments summarized above and found that PR was at an acceptable levels when PR<5 mm h1 whereas the tipping bucket measurements over a 1 min interval can be in error of about 12 mm h1. They also suggested that the relative error in PR values can reach to +20% for heavy rain, 50% for average rain episodes and +300% for light rain and drizzle. Overall, their results suggested that uncertainties were about 1214% for PR<5 mm h1 and 3840% for PR<5 mm h1. Disdrometer measurements can have large uncertainty at low PR values because of their small sampling area (50 cm2) and a lower particle threshold size about 300500 μm . The PR and PA cannot be obtained accurately when the droplet sizes are less than 500 μm (0.5 mm). It is important to note that the duration of precipitation with small PR and PA values can be much higher than those of high PR events. But, if PR=1 mm h1 occurs over a 10 h time period then it can result in PA=10 mm which can not be measured accurately by various instruments. The standard procedure for intercomparisons among various precipitation measurements obtained from the instruments is usually done against a standard one e.g. a manual or weighing gauge but, for a light precipitation, this method doesnt work accurately; therefore, relative intercomparisons in this study are made (e.g. instrument to instrument). The measurements collected during the Fog remote sensing and modeling (FRAM) projects (Gultepe et al. 2007) will be summarized to better understand (1) light precipitation events that include heavy fog and drizzle, (2) precipitation rate accuracy and (3) drizzle effect on visibility.