NetSTORM Precipitation Frequency Statistics--A Complementary Tool to NOAA's Atlas 14 and a Data Processing Powerhouse

OVERVIEW A major US water utility director was recently quoted saying that 'Every 10 years or so we're having a 100-year storm' (Boston Globe, July 23, 2021). While this statement reflects a reasonable perception that the frequency of intense rainfall is increasing across much of the United States and worldwide, it misleadingly suggests that existing precipitation frequency estimates have little value, potentially sowing distrust in engineering methods used to assess infrastructure performance. The precipitation frequency statistics that steer collection system assessment and design should be transparent, defensible, and accessible. CDM Smith's NetSTORM software (Heineman, 2004) and companion precipitation statistics website (http://dynsystem.com/netstorm/IDFStats.htm) provide data and analysis tools to efficiently analyze rainfall data and develop intensity and duration statistics to compare with those published by agencies such as the National Oceanographic and Atmospheric Administration (NOAA). Because rainfall intensity varies spatially and temporally, a single storm may cause extreme rainfall at one location over a certain duration, lesser rainfall at another location over a different duration, and yield ordinary accumulations elsewhere. In assessing rainfall across a metropolitan area, it is thus reasonable to expect numerous '100-year storms' over a decade. The readily available and up-to-date statistics for principal rain gages that NetSTORM offers can help dispel mistrust of rainfall frequency statistics based on historical data. This paper will present the existing industry standard source of precipitation statistics and its limitations, then discuss how NetSTORM can lift those constraints as well as streamline the process of general rainfall data acquisition, QAQC, and analysis. LIMITATIONS OF THE CURRENT INDUSTRY STANDARD PRECIPITATION STATISTICS NOAA's Atlas 14 precipitation frequency estimates (Perica et al., 2018) serve as standards for a wide variety of design and planning activities and regulations across the US. Its Precipitation Frequency Data Server (PFDS) provides online access to the frequency estimates presented in the atlas. However, the PFDS is lacking in several areas: 1. NOAA PFDS datasets are only as recent as the data used to develop Atlas 14, which has no data after 2000 for the 17 states analyzed in its first two volumes, and lacks coverage for five northwestern states, 2. The NOAA PFDS provides statistics for only a select set of durations and average recurrence intervals (ARI), limiting a user's ability to easily quantify ARIs outside of or between those values; and 3. While NOAA's PFDS provides the information necessary to evaluate the significance of historical rainstorms, it does not provide software tools for quickly and easily applying its statistics to timeseries data. NETSTORM: A COMPLEMENTARY TOOL TO ATLAS 14 AND A DATA PROCESSING POWERHOUSE Over the course of its 30-year history, NetSTORM has been continuously refined to meet these needs of (1) providing statistics on the latest data, (2) identifying storms with an ARI of less than one year, and (3) streamlining the process of precipitation data acquisition, QAQC, and analysis. The NetSTORM Precipitation Frequency Statistics website includes up-to-date statistics for 100 cities across the United States. The free NetSTORM software package can produce statistics from any dataset available to a user and provides a wider range of frequency estimates than the NOAA PFDS. NetSTORM can estimate storm ARIs on a continuous scale that extends below one year (the lower limit of the NOAA PFDS). Additionally, ARIs estimated by NetSTORM are presented alongside interpolated ARIs from Atlas 14. Given a lengthy rainfall dataset with thousands of records, any NetSTORM user can quickly produce a rich and portable html report. Its rainfall statistics reports provide tabular and graphic displays for each major historic storm at various durations, giving users easy access to frequency statistics for individual storms. Annual maximum series graphs (Figure 1) allow users to visually assess heavy rainfall frequency across historic records. CASE STUDY: RAPID ANALYSIS AND CONTEXTUALIZATION OF TROPICAL STORM FRED AND HURRICANE IDA NetSTORM was used to rapidly contextualize the impacts from both Tropical Storm Fred (8/19/2021) as well as Hurricane Ida (9/1/2021) for both Hartford's Metropolitan District (MDC) and the Town of West Hartford. Both storms caused significant flooding in the Hartford metropolitan area. Immediately after both storms, the latest rainfall data from multiple local rain gages in the area were consolidated and processed with NetSTORM, identifying rainfall and ARIs at various durations. The data output from NetSTORM were further processed to produce subplots comparing the hyetographs, peak hourly intensities, and maximum ARIs with those of Fred and Ida (Figure 2). NetSTORM's calculations show how the two storms over 24 hours are similar, averaging about 5.25 inches, which amounts to an ARI of about 12 years. However, Fred was actually a much rarer event, with a peak intensity exceeding 3 inches per hour and a peak ARI approaching 300 years based on 4.76 inches of rain over its most intense 3 hours. NetSTORM enables these kinds of rapid analyses that better contextualize real world conditions with historical data, a service that is becoming ever more critical with evolving global climate patterns in an increasingly data-driven regulatory landscape. CONCLUSION Precipitation frequency estimates presented on the NetSTORM website use methods that are largely compatible with Atlas 14. NetSTORM is available for download from the website (dynsystem.com/netstorm), allowing users to perform comparable calculations for other locations or develop custom analyses. The website and the software can help inform public understanding of extreme rainfall frequency and streamline the otherwise arduous process of precipitation timeseries data acquisition, QAQC, and analysis. While non-stationarity of rainfall now and in the future means that precipitation frequency estimates are subject to change, the statistics published in Atlas 14 and elsewhere are generally valid for understanding historic rainfall, and likely only warrant modest adjustment for assessing extreme rainfall over the next few decades. NetSTORM is an ideal choice for rainfall data processing and analysis, providing a concise interface, practical and portable report, and extended statistics when compared with Atlas 14.