PI: Shang-Ping Xie, Co-PI: Y. Chen and J. Hafner

IPRC/SOEST Univ. of Hawaii at Manoa

The project was motivated by a need of high resolution climate data for Hawaiian Islands. With respect to climate data it seems that Pacific region including Hawaii is not covered in sufficient details. Mainland US states enjoy better representation of long term climate data at higher spatial and temporal resolution, which makes various research studies plausible. However, it is not the lack of data per se that is impeding the progress in local and/or small scale climate studies and research in Hawaii, but rather it is the accessibility of the data. In principle the data are there, but scattered in various data centers and stored in incompatible formats. Thus, the primary goal of this project is to collect, process and remapped relevant data and present to general public through one data server. In particular, the project focuses firstly on processing of high resolution radar and satellite data, secondly it demonstrates the use of the data in scientific research application. The research application was aimed to answer the puzzling question about unusual winter 2004/2005 rainfall (more details will follow).

  1. Data

The following data are considered : NEXRAD (Next Generation Weather Radar), MODIS (Moderate Resolution Imaging Spectroradiometer), LIS (Lightning Imaging Sensor) and in situ data.

      1.1 NEXRAD

      NEXRAD data are archived at NCDC site and in Hawaii there are 4 radar sites: South Kauai, Molokai, Kamuela and South shore of Hawaii (map). Three winter months (Nov. Dec. and Jan.) 2004/2005 were selected. Variables surface rain rate, base reflectivity at 124 and 248 Nm range, top echo and vertically integrated liquid water were selected for merging and remapping on a regular lat/lon grid with 1 km grid spacing. The 1 km grid box size was chosen to be close to the original radar data pixel size which is 0.54 Nm x 1 angle degree (close to 1 km). Upon transferring the selected data from NCDC they were converted into binary format, remapped and merged on one map. The resulting averaged fields are shown in the following table. Variables shown here are 3-month diurnal averages.


rain rate (mm/hr)

base reflectivity (dBZ)

vertical liquid water (cm)

animated gif movie

average 24hr. loop

average 24hr. loop

average 24hr. loop

      1.2 MODIS

      MODIS data were obtained from NASA's DAAC for 3 winter months 2004/2005. The cloud mask and SST MODIS products were considered along with the geolocation data necessary for remapping. The MODIS cloud products recognizes 4 cloud categories: confident clear, probably clear, uncertain and confident cloudy. The resulting cloud coverages presented here are based on confident cloudy category. The daytime and nighttime cloudiness and clear sky SST are given below. MODIS clouds

      1.3 LIS

      LIS (Lightning Imaging Sensor) data were obtained from MSFC NASA archive and converted into GrADS binary format and on a 0.5 degree lat/lon map. The winter 2004/2005 (Nov. Dec. and Jan.) average is given in the figure below. LIS winter

      1.4 Station data

      Station data were acquired from NCDC archive and the following datasets were converted into GrADS station format: hourly rainfall (1950-2005), hourly station data (1998-2005) and rawindsonde data (1998-2005). Rainfall station, hourly weather and sounding data plots are given below.

      rainfall nov04-jan05
      Averaged monthly rainfall (mm) Nov.-Dec. 2004 + Jan 2005, left Kauai, right Oahu

      rainfall nov04-jan05
      Averaged monthly rainfall (mm) Nov.-Dec. 2004 + Jan 2005, left Maui, right island of Hawaii

      Hourly station data as the winter average is given in the figure below:

      Sounding data were collected for stations Lihue and Hilo starting 1998 until April 2005. The averaged sounding data for each station for winter 2004/2005 are given below. Wind speed is in color and direction in bars, temperature in contours.

      Hilo wind profile

      Lihue wind profile

      An example of data integration is given below, where averaged MODIS derived cloudiness is compared to corresponding time period of TMI cloud liquid water content.

      MODIS clouds day+night
      TRMM CLW(mm)