JP1.2 THE KA`Ū STORM: IMAGING PRECIPITABLE WATER USING GPS James Foster*, Michael Bevis, Steven Businger and Yi-Leng Chen University of Hawaii, Honolulu, HI 1. INTRODUCTION The Ka`ū (pronounced Ka-oo) Storm was an extreme rain event that impacted the south and east sides of the Big Island of Hawai`i during the 1 st and 2 nd of November, 2000. Maximum hourly rain-rates were over 100 mm/hr (~4 in/hr) and the total rainfall from the storm reached nearly 1000 mm at one location, with a 24-hour accumulation that fell just short of setting a state record. Stream gauge records show that this was the most intense, widespread rain event in 20 years, and at several sites the maximum stream-flow from this storm established records. The extensive flash floods that resulted are estimated to have caused $70M property damage and the impacts on roads and other infrastructure persisted for weeks afterwards. Intriguingly, several days after this huge water-load was deposited on the island GPS sites recorded the first documented instance of an aseismic slip event on Kīlauea Volcano (Cervelli et al. 2002). This suggests that a shallow fault may have been activated by the increased pore-pressure due to this excess of water, and raises the question of whether landslides might also be a delayed but significant hazard from extreme rainfall events in this area. The maximum rainfall was recorded by the rain gauge at Kapāpala Ranch on the south slope of Mauna Loa (Figure 1). This gauge recorded 989 mm (>39 inches) of rain over 36 hours and heavy rainfall was recorded over most of the southern and eastern portions of the Island. Previous studies have investigated this pattern of heavy rain event on the south-facing slopes of Hawai`i (Kodama and Barnes, 1997, Schroeder, 1978) and have interpreted the large temporal scales of the heavy precipitation and the quasi-stationary nature as the effect of orographic lifting which anchors the storm to the mountainous terrain when the winds are blowing onshore. * Corresponding author address: James Foster, POST 602, 1680 East West Road, 1680 East West Road, Honolulu, HI 96822; e-mail jfoster@soest.hawaii.edu The formation of the deep convection required for heavy rain is controlled by the depth of the moist layer depth and the interaction of the surface winds with the islands’ complex topography. The NCEP Global Spectral Model, which, with about 1-degree grid spacing, does not include topography for the Hawaiian Islands, failed to predict any of the subsequent rainfall. The more detailed Regional Spectral Model (10- km grid) run operationally by NWS/U.H. Dept. Meteorology (Wang et al. 1998) which incorporates coarse topography was able to predict some rainfall, but it was both grossly underestimated, and mislocated. Only with the inclusion of detailed topography in a 3-km grid Mesoscale Spectral Model (Zhang et al. 2000; http://www.soest.hawaii.edu/~rsm), run after the event, was a more realistic model of the actual rainfall generated although even then the predicted rainfall was only 50-60% of the actual measured values. This highlights the potential contribution that an all-weather GPS system could make in providing rapid and accurate precipitable water (PW) estimates, particularly in concert with satellite data, for which GPS could provide point PW calibrations (Motell et al., 2000). A network of GPS receivers (Fig. 1) covers much of Kīlauea and the summit of Mauna Loa, allowing us to calculate frequent, accurate estimates for the precipitable water over much of the area most affected by the storm. With sites at elevations from sea level to over 4000 m at the summit, and with an average spacing of less than 10 km this network provides us with a unique opportunity to examine the details of the precipitable water distribution as the storm passed over the southeastern section of the island. 2. DATA AND METHODOLOGY The Big Island GPS network consists of 7 receivers run by the Pacific GPS Facility, 15 more operated by the Hawaiian Volcano Observatory (HVO-USGS) and Stanford University, and one final receiver (MKEA) available from the International GPS Service (IGS). The rain gauge data is compiled from the National Weather Service Hydronet network with