• SCIENTIFIC BACKGROUND AND RATIONALE
• ARRAY DESIGN
  • Instrumentation
  • Station Locations and Implementation Schedule
• DATA DISPLAY
• REFERENCES

In the atmosphere, release of latent heat by precipitation is the primary source of energy for driving the atmospheric general circulation. Rainfall also represents an important source of buoyancy to the upper ocean, affecting horizontal and vertical distribution of mass, the oceanic circulation, and the intensity of turbulent vertical mixing. Rainfall related physical processes are particularly pronounced in and over the tropical oceans, where over 2/3 of the global precipitation falls. The tropical Pacific is, moreover, characterized by the highest rain rates in the world ocean (up to 5000 mm per year), by significant rainfall variations associated with the seasonal migration and intensification of major convergence zones (e.g., the Intertropical Convergence Zone and the South Pacific Convergence Zone), and by major interannual perturbations of the climate system associated with ENSO (Janowiak and Arkin, 1991).

Despite its significance for understanding oceanic variability, atmospheric variability and coupled ocean-atmosphere interactions in the tropics, rainfall fields are extremely difficult to determine with accuracy suitable for climate purposes. Island rainfall measurements (i.e., Morrissey et al., 1995) can be very valuable sources of rainfall information; unfortunately, islands are not uniformly distributed in climatically important regions of the world ocean, and in not all cases are island measurements representative of open ocean conditions because of potential island heating and sea breeze effects. Present weather reports from volunteer observing ships (e.g., Dorman and Bourke, 1979), though quasi-global in coverage, are sparsely distributed in space and time, and qualitative by their very nature (Petty, 1995).

Limitations in conventional oceanic rainfall observations have long stimulated interest in obtaining estimates of precipitation from satellite, and several techniques now exist based on infrared and microwave retrievals (Arkin and Ardunay, 1989). The measurements derive primarily from operational polar orbiting and geostationary satellites launched for weather and defense applications. To these ongoing operational measurements will soon be added data from TRMM, a joint NASA/ NASDA research effort aimed primarily at the documentation and understanding of rainfall variability in the tropics, and its impact on the climate system (Simpson et al, 1988). TRMM, launched in November 1997, isinstrumented with passive microwave a as an active precipitation radar. Orbital inclination is 35° at an altitude of 350 km. TRMM's primary measurement objective is to provide monthly mean precipitation rates on a 5° by 5° basis in tropical regions with a relative accuracy of 10%.

TRMM data are greatly stimulating progress in tropical oceanography, meteorology, and hydrology. Nonetheless, like other existing satellite rainfall measurements, those from TRMM are affected by random sampling error, algorithm errors and biases due to the indirect nature of rainfall estimates. Special validation efforts for TRMM will address some of the uncertainties in calibrating and validating the satellite estimates. However, these efforts are of limited geographical extent and temporal duration in oceanic settings (e.g., at Kwajalein atoll for several weeks in 1999).

The TAO/TRMM rainfall array builds on the existing TAO Array (McPhaden, 1995) in the equatorial Pacific. It spans nearly the full zonal extent of the basin between 8°N and 8°S. Moored rainfall time series are collected in several climatic regimes in the tropical Pacific, namely, the Intertropical Convergence Zone, the equatorial cold tongue, the western Pacific warm pool and the South Pacific Convergence Zone. There are broad geographical overlap with island measurements in the western Pacific (Morrissey et al., 1995); however, the array extends into the central and eastern equatorial Pacific which is essentially devoid of islands. The meridional section along 140°W samples a region of particular interest, namely, the ITCZ of the eastern Pacific where significant discrepancies between satellite and conventional rainfall estimates have been found by Spencer (1993) and Janowiak et al. (1995). The line along 95°W samples a region of strong monsoonal wind forcing and sharp meridional contrasts in climatological SST, cloudiness, surface heat fluxes and rainfall. Ocean-atmosphere-land interactions result in a pronounced seasonal cycle in this region, with implications for rainfall variability both over the ocean and the adjoining land masses (Mitchell and Wallace, 1992).

This program provides extensive open-ocean ground based support for TRMM in the form of in situ rainfall estimates in the tropical Pacific to help reduce the uncertainty of TRMM (and other) spaceborne estimates of rainfall. The time series data are also valuable for characterizing the time and space scales of rainfall variability in the tropics, for documenting the impact of rainfall on the ocean, and for better understanding the role of the hydrologic cycle in the climate system.

ARRAY DESIGN

Instrumentation
The array consists of a subset of existing TAO Next Generation ATLAS moorings instrumented with RM Young capacitance rain gauges. The capacitance gauge works on the principal of collecting rainfall in a cylindrical tube and measuring the amount of water accumulated at fixed intervals. The time difference of accumulations thus gives a measure of rain rate. Tests of the RM Young rain gauges both in the equatorial Pacific and in the Olympic Rainforest of Washington State have lead us to conclude that they are the best commercially available product for long term moored applications. RM Young capacitance gauges were also used on the WHOI IMET mooring and on various ships during TOGA-COARE.

Rain accumulations will be measured at 1-minute intervals, from which first differences will be taken to compute rain rates. Resolution for rain accumulation is 0.004 mm, and the manufacturer's stated accuracy for accumulation is 1 mm. Laboratory calibrations of the RM Yound siphon gauge at PMEL, however, typically show maximum residuals about calibration curves of 0.3 mm, significantly lower than the manufacturer's specification.

The RM Young may be subject to a wind speed bias which, uncorrected, could lead to underestimates at the high end (> 10 m/s) of wind speeds observed in the tropical Pacific. This bias can be corrected though using simultaneous TAO wind measurements and a simple algorithm for catchment type rain gauges (Bradley and Paulson, 1996).

Each TRMM rainfall site is instrumented with a surface conductivity (i.e. salinity) sensor. Conductivity is measured from a Sea-Bird SEACAT type sensor. Subsurface conductivity measurements at a few selected sites will be considered as funding permits.

Station Locations and Implementation Schedule

The array will consist of 28 sites when fully implemented (solid blue and red symbols in figure at right). Support for rainfall measurements along 95°W is provided by NOAA's EPIC program.

Horizontal correlation between rainfall and salinity at adjacent stations is likely to be significant on time scales of days over the 2-3° array separations in the meridional direction and weeks over the 10-15° separations in the zonal direction based on analysis of COARE data (Thiele et al, 1995). Hence, the array will provide a coherent picture of rainfall variations for monthly and longer time scales of principal climatic interest.

The proposed schedule for deployment is such that the array, which was comprised of 10 sites at the end of 1998, will be fully implemented by late 2000. This array will be complemented by TRITON moorings which will also measure rainfall using optical raingauges along 156°E, 147°E, and 138°E.

DATA DISPLAY

TAO Enhanced Sensor Display. Data are available for delivery on the TAO/TRITON delivery page.

Rainrate data will be quality controlled at PMEL, and distributed via the World Wide Web after an initial period of evaluation. Yearly submissions will be made to the National Climatic Data Center and the Global Precipitation Climatology Center.