An Assessment of Automated Rain Intensity Correction in the RainVue™ Gauge: Part 1. Field Setup and Daily Comparisons

Beginning in mid-April 2023, a comparison among five rain gauges was initiated at a field site in Norman, a comparison that is ongoing. The primary purpose of the field study was to compare observations from two RainVue™ tipping-bucket gauges (Campbell Scientific, Inc.) to observations from two 380C tipping-bucket gauges (Met One Instruments, Inc.) and a T-200B vibrating-wire weighing gauge (Geonor, Inc.). The RainVue gauges are unique in that they have a microprocessor to account for lack of accumulation during time of tip, a familiar problem with tipping-bucket gauges, especially during high intensity rain events. The RainVue instruments also have an aerodynamic mechanical design to reduce undercatch due to wind.

Central to the field observations is a pit in which one of the 380C gauges and the T-200B gauge are located such that the orifice of each is at ground level, thus minimizing the influence of wind. The two RainVue gauges were located above ground on the pit grill in a south-north line separated by 1.8 m from mid-April to mid-October 2023 after which one of the gauges was placed in the pit with its orifice at ground level, similar to the two other gauges in the pit. The second 380C is located 100 m from the pit and is a component of the Oklahoma Mesonet station (NRMN) in Norman. Photos of the field setup will be shown.

There are two measurement periods of interest. During Period 1, both RainVue gauges were aboveground and one-minute accumulation data were collected for 55 rain days from mid-April to mid-October. During Period 2, one of the RainVue gauges was in the pit and one-minute data were collected for 74 rain days from mid-October 2023 through June 2024. A rain day is defined as any amount of rain recorded by the T-200B during a UTC Day. There are numerous components to a complete comparative data analysis of the five gauges. In the interest of limited space and time, we highlight a few important results from Periods 1 and 2 that involve only the gauges located on or in the pit. In this Part 1 of 2, we will show overall results in accumulation from the two periods while Part 2 will deal with rain intensity comparisons. Various graphs will be shown to explain the numerical results.

Period 1

For simplicity, we refer to the south RainVue gauge as SRV and the north RainVue gauge as NRV, which is in the pit during Period 2. Over the 55 days when both RV gauges were aboveground, the SRV gauge recorded 673 mm and NRV gauge 679 mm, a difference of 0.9%. Of the 55 days, SRV was greater than NRV on 17 days, lesser than NRV on 16 days, and zero difference on 22 days, indicating no bias. If we ignore the effect of the microprocessor and simply multiply the number of tips by 0.254 mm/tip, the difference for both gauges is 1.0%, that is, the 55-day totals cited above would be reduced by approximately 6.8 mm. While 1.0% is a small difference, among all rain days, the maximum daily difference is 4.3% and 4.6%, respectively, for the SRV and NRV gauges. This demonstrates that there are many days where corrections for rain intensity are negligible while on days with high rain intensity the correction can be substantial.

Period 2

During the 74 rain days when the NRV gauge was in the pit for comparison with the aboveground SRV and the collocated T-200B and the 380C gauges. The four gauges observed 551 mm (NRV), 523 mm (SRV), 561 mm (T-200B), and 555 mm (380C), respectively, over the 74 rain days. The T-200B is used as the reference gauge because, while all three gauges in the pit experience minimal wind effects, a weighing gauge is expected to be much less vulnerable to error due to intensity than a tipping bucket due to mechanical differences. We see that the SRV gauge observed 5.1% less rain than the NRV gauge. Considering that the NRV gauge rain total for Period 1 was about 1% greater than that from the SRV gauge, this suggests that the wind effect on the aboveground SRV was about 4%. Relative to the T-200B gauge, the NRV was 1.8% less. If we ignore the effect of the microprocessor and simply multiply the number of tips by 0.254 mm/tip, the difference for both gauges is approximately 0.4%, that is, the 74-day totals cited above would be reduced by approximately 2.2 mm and 2.5 mm, respectively, for the SRV and NRV gauges. Among all rain days, the maximum daily difference is 3.4% and 3.3%, respectively, for SRV and NRV gauges. As in Period 1, we conclude that the rain intensity correction will have its greatest impact during high intensity rain events.

We observe that the 74-rain day accumulation from the 380C gauge lies roughly mid-way between the T-200B and NRV. Thus, with the T-200B as the reference, both the 380C and NRV gauges exhibit undercatch related to time-of-tip. Fortunately, the opportunity to improve the estimation of rainfall is available with RainVue gauges.