• Anderson, S. P., and M. F. Baumgartner, 1997: Radiative heating errors in naturally ventilated air temperature measurements made from buoys. J. Atmos. Oceanic Technol.,15, 157–173.

    • Crossref
    • Export Citation
  • ——, R. A. Weller, and R. B. Lukas, 1996: Surface buoyancy forcing and the mixed layer of the western Pacific warm pool: Observations and 1D model results. J. Climate,9, 3056–3085.

    • Crossref
    • Export Citation
  • Buck, A. L., 1981: New equations for computing vapor pressure and enhancement factor. J. Appl. Meteor.,20, 1527–1532.

    • Crossref
    • Export Citation
  • Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S. Young, 1996: Bulk parameterization of air–sea fluxes for TOGA COARE. J. Geophys. Res.,101 (C2), 3747–3764.

    • Crossref
    • Export Citation
  • Flament, P., and M. Sawyer, 1995: Observations of the effect of rain temperature on the surface heat flux in the intertropical convergence zone. J. Phys. Oceanogr.,25, 413–419.

    • Crossref
    • Export Citation
  • Gill, G. C., 1983: Comparison testing of selected naturally ventilated solar radiation shields. NOAA Data Buoy Office Report, Contract NA-82-0A-A-266, NOAA/National Data Buoy Center, Bay St. Louis, MS, 15 pp.

  • Gosnell, R., C. W. Fairall, and P. J. Webster, 1995: The sensible heat of rainfall in the tropical ocean. J. Geophys. Res.,100 (C9), 18 437–18 442.

    • Crossref
    • Export Citation
  • Hosom, D. S., R. A. Weller, R. E. Payne, and K. E. Prada, 1995: The IMET (Improved Meteorology) ship and buoy system. J. Atmos. Oceanic Technol.,12, 527–540.

    • Crossref
    • Export Citation
  • Kincaid, D. C., and T. S. Longley, 1989: A water droplet evaporation and temperature model. Trans. ASAE,32, 457–463.

    • Crossref
    • Export Citation
  • Kinzer, B. D., and R. Gunn, 1951: The evaporation, temperature, and thermal relaxation-time of freely falling waterdrops. J. Meteor.,8, 71–83.

  • List, R. J., 1984: Smithsonian Meteorological Tables. Smithsonian Institution Press, 527 pp.

  • Nystuen, J. A., J. R. Proni, P. G. Black, and J. C. Wilkerson, 1996: A comparison of automatic rain gauges. J. Atmos. Oceanic Technol.,13, 62–73.

    • Crossref
    • Export Citation
  • Webster, P. J., and R. Lukas, 1992: TOGA COARE: The Coupled Ocean–Atmosphere Response Experiment. Bull. Amer. Meteor. Soc.,73, 1377–1416.

    • Crossref
    • Export Citation
  • Weller, R. A., and S. P. Anderson, 1996: Surface meteorology and air–sea fluxes in the western equatorial Pacific warm pool during TOGA Coupled Ocean–Atmosphere Response Experiment. J. Climate,9, 1959–1990.

    • Crossref
    • Export Citation
  • View in gallery

    Diagram of the precipitation temperature module. It indicates the location of the four thermistors: PT1, PT2, PT3, and PT4.

  • View in gallery

    Observed precipitation. The observed rain rate (black line) and the cumulative rainfall (gray line) are shown for the two deployments. The data represent 1-min averages.

  • View in gallery

    Rain rate and droplet temperatures for event 3. Data shown are for 5.5 h during 28 October 1992. (a) The observed rain rate (black line) and the cumulative rainfall (gray line) are shown for the two deployments. (b) The four temperatures from the PTM are plotted for the corresponding time period (thin black lines). Also shown are the sea surface (dashed), dry-bulb (thick black), and wet-bulb (thick gray) temperatures. The data represent 1-min averages.

  • View in gallery

    Rain rate and droplet temperatures for event 4. Same as Fig. 3 except data are from 5 h on 13 December 1992.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 316 61 10
PDF Downloads 156 80 5

Moored Observations of Precipitation Temperature

View More View Less
  • 1 Woods Hole Oceanographic Institution, Woods Hole, Massachusetts
Full access

Abstract

Direct observations of precipitation temperature were made from a surface buoy deployed for four months in the western Pacific warm pool. The observed rain droplet temperatures are equal to the wet-bulb temperature to within the measured wet-bulb temperature uncertainty of ±0.4°C. The rain droplet temperatures are 4.8°–5.8°C cooler than the ocean surface temperature. The sensible heat flux associated with the rain is found to be a significant component for the net surface heat while it is raining, ranging from −65.0 to −204 W m−2 (ocean cooling) and accounting for 15%–60% of the net heat flux for any single rain event. The rain heat flux is also important on longer timescales in the warm pool, where there is a close balance between surface heating and cooling and high precipitation rates. During the 4-month deployment period, the rain heat flux is 2.8 W m−2 (ocean cooling) and 15% of the net surface heat flux.

Corresponding author address: Dr. Steven P. Anderson, Physical Oceanography Department, Woods Hole Oceanographic Institution, MS 29, Woods Hole, MA 02543.

Abstract

Direct observations of precipitation temperature were made from a surface buoy deployed for four months in the western Pacific warm pool. The observed rain droplet temperatures are equal to the wet-bulb temperature to within the measured wet-bulb temperature uncertainty of ±0.4°C. The rain droplet temperatures are 4.8°–5.8°C cooler than the ocean surface temperature. The sensible heat flux associated with the rain is found to be a significant component for the net surface heat while it is raining, ranging from −65.0 to −204 W m−2 (ocean cooling) and accounting for 15%–60% of the net heat flux for any single rain event. The rain heat flux is also important on longer timescales in the warm pool, where there is a close balance between surface heating and cooling and high precipitation rates. During the 4-month deployment period, the rain heat flux is 2.8 W m−2 (ocean cooling) and 15% of the net surface heat flux.

Corresponding author address: Dr. Steven P. Anderson, Physical Oceanography Department, Woods Hole Oceanographic Institution, MS 29, Woods Hole, MA 02543.

1. Introduction

Although precipitation is known to play an important role in the buoyancy flux at the ocean surface by freshening the ocean, the heat flux associated with the precipitation is usually ignored. When the rain droplet temperature is different than the ocean surface temperature, there is a sensible heat flux associated with the mass flux. Models and laboratory studies of the evaporation and temperature of a freely falling droplet of water suggest that the temperature of rain will be near the ambient wet-bulb temperature (Kinzer and Bunn 1951; Kincaid and Longley 1989; Gosnell et al. 1995). Observations made in the intertropical convergence zone show that the sensible heat flux associated with rain may account for 40% of the net surface heat flux while it is raining (Flament and Sawyer 1995).

The theoretical and experimental work of Kinzer and Gunn (1951) suggests that a free-falling water droplet will reach an equilibrium temperature that is slightly lower than the ambient wet-bulb temperature. Kincaid and Longley (1989) state that the amplitude of this difference is determined by the empirical function of the diffusivity of water vapor in air used in the model. They predict droplet temperatures of 0.2°–0.4°C below the wet-bulb temperature using the diffusivity function given by List (1984). They suggest that a water droplet evaporates and cools more efficiently than a wick-covered thermometer and that the droplet temperature may be closer to the true wet-bulb temperature than the measured temperature.

Gosnell et al. (1995) use a microphysical model of rain falling through the atmospheric boundary layer and suggest that due to the finite thermal relaxation time, large raindrops will be cooler than small drops and may reach the surface at temperatures up to 0.2°C below the ambient surface wet-bulb temperature. The rate at which a freely falling water droplet reaches equilibrium at the wet-bulb temperature is a function of both the size of the droplet and the speed through the air. This adjustment takes several seconds. As the droplet falls through the atmosphere, it encounters air of increasingly higher temperature determined by the adiabatic temperature profile. Different convective structures in the atmosphere will have different adiabatic profiles and droplet sizes that both determine the temperature difference.

The demand for increased accuracy in the estimates of air–sea fluxes led to the development of an instrument to measure rain droplet temperature. This instrument was designed to be deployed on a surface mooring along with the traditional surface meteorological and radiation instrumentation. This paper describes the precipitation temperature module (PTM) that was developed for the Improved Meteorological Package (IMET) (Hosum et al. 1995) and presents observations of precipitation temperature taken during the Tropical Ocean Global Atmosphere (TOGA) Coupled Ocean–Atmosphere Response Experiment (COARE). COARE took place in the western equatorial Pacific Ocean in a region where annual average upper-ocean temperatures exceed 29°C (Webster and Lukas 1992). A complete description of the mooring on which the IMET and the PTM were deployed and of the surface meteorological observations is given by Weller and Anderson (1996).

2. Instrumentation

a. Precipitation temperature

The current design of the PTM is based on experiments conducted at the Woods Hole Oceanographic Institution (WHOI) in the fall of 1991. During these experiments a funnel from an R. M. Young model 50202 precipitation gauge was modified with temperature sensors arranged as in Fig. 1, along with a temperature sensor located near the rain gauge and directly in the flow path. The sensor, located directly in the rain’s path, is continually being flushed with precipitation thermally uncontaminated by the funnel and, therefore, provides a close approximation of the precipitation’s temperature. Results of this experiment showed that when a heavy rain event occurred the sensor located directly in the rain reached a steady-state value and eventually all the sensors reached this same equilibrium temperature, indicating that the funnel was at (or very close to) the temperature of the precipitation. These results indicate that a funnel with temperature sensors, as distributed in Fig. 1 to measure the thermal state of the funnel, could be used to measure the temperature of precipitation during heavy precipitation events.

To improve the response time of the PTM, the R. M. Young funnel was replaced with a thin stainless steel funnel with less thermal mass (Fig. 1). The stainless steel funnel quickly comes to equilibrium temperature as the rainwater flows through and the funnel temperature is sampled. By using a lower thermal mass funnel, the response time of the funnel is increased so that rapid changes in the precipitation temperature can be measured. The funnel was made of a 0.003-in.-thick stainless steel sheet and has a sampling area of 100 cm2. At the base of the funnel is a 1-cm-diameter drip mouth that allows the precipitation to flow freely out the bottom. Four independent temperature measurements are made using fast-response thermistors in a linear bridge configuration. The measurement range is 0°–38°C, with 0.1°C accuracy and 0.037°C resolution. Thermistor 1 (PT1) is suspended in the center of the funnel, thermistor 2 (PT2) is mounted halfway down the neck of the funnel, thermistor 3 (PT3) is located at the mouth of the funnel, and thermistor 4 (PT4) is located in the funnel outflow below the mouth (Fig. 1). The thermistors are sampled at 0.1 Hz and averaged over 1 min before being recorded by the IMET data logger.

Electronics of the PTM consisted of the standard IMET power and communications and sensor control module boards. A special low-power analog front-end board was designed to measure four independent channels of temperature with 10 bits of resolution. The analog front end was interfaced to a standard IMET module, and custom firmware was written so the completed PTM module responded on serial communication as a standard IMET module. By using the IMET boards, the PTM was easily integrated into the suite of other IMET sensors on the buoy. Upon interrogation of the PTM by the IMET data logger, the PTM responds with four temperature measurements averaged over the previous minute.

b. Wet-bulb temperature

Humidity is measured with the IMET relative humidity module that uses a Rotronics relative humidity sensor (model MP-100) and a platinum resistance thermometer. The sensors are mounted inside a multiplate radiation shield (R. M. Young Model 41002) based on the design by Gill (1983). The humidity sensor is mounted inside a porous Teflon shield that allows vapor penetration but protects the sensor from salt spray and rain. Laboratory calibrations suggested accuracies of 2%–3% RH at humidities of up to 95% RH. The instruments are accurate to 3%–5% RH over the four-month TOGA COARE deployment in the field (Weller and Anderson 1996). The platinum resistance thermometer, which measures the temperature in the proximity of the Rotronics relative humidity sensor, has an accuracy of better than 0.01°C. In the case of low wind speeds and high solar insolation, the temperatures in the unit can be as high a 2°C above ambient air temperature. In this case, the measured relative humidity is less than the ambient humidity due to the increased temperature. However, the vapor pressure inside the Gortex cover is theoretically equal to the ambient vapor pressure and can be estimated by using the internal temperature measurement.

The relativity humidity, U, is equal to the ratio of the mixing ratio, r, to the saturation mixing ratio, rw. The mixing ratio is determined by
i1520-0426-15-4-979-e1
where P is the atmospheric pressure, e′ is the vapor pressure, and ɛ is a constant equal to 0.62197. The vapor pressure is then given by
i1520-0426-15-4-979-e2
The saturation vapor pressure, ew, is calculated using bulk formulas (Buck 1981):
i1520-0426-15-4-979-e3
where P is the atmospheric pressure (mb), T is the dry-bulb temperature (°C), and ew is given in millibars. The wet-bulb temperature is the temperature that a parcel of air, originally at a given T, P, and r, assumes when water is slowly added to it at the parcel’s current temperature and evaporated into that air parcel adiabatically at constant pressure until the parcel reaches saturation. The wet-bulb temperature, Twet, is related to T, P, and r, using Ferrel’s psychometric equation (from List 1984, 365)
i1520-0426-15-4-979-e4
where ewet is the saturation vapor pressure at Twet. Numerical iteration is used to solve (4) for Twet at a given observed temperature, pressure, and humidity.

The largest error in the wet-bulb temperature comes from the uncertainty in the relative humidity measurement. The relative humidities, and the associated uncertainties, are generally high during rain events. As an example, assume an air temperature of 25°C, a barometric pressure of 1000 mb, and a humidity of 95 ± 3% RH. Propagating the humidity uncertainty through the calculation, the estimated wet-bulb temperature is 24.2 ± 0.4°C. In general, the wet-bulb temperatures calculated this way have an accuracy of only ±0.4°C.

c. Dry-bulb temperature

The dry-bulb temperature is measured with the IMET air temperature module, which uses a platinum resistance thermometer shielded by a multiplate radiation shield. The thermistor is sampled every 6 s, and 60-s averages are recorded. The air temperatures are accurate in the field to 0.1°C during the nighttime. During the day, the solar radiative errors may be as large as 1.5°C during periods of low winds (<2 m s−1) and high solar insolation (Anderson and Baumgartner 1997). The precipitation events presented in this paper occur for the most part during the night or under cloud cover, so the solar radiative errors are not of consequence.

d. Precipitation

A siphon rain gauge (R. M. Young Model 50202) was used on the buoy to measure rain rate. This is a volumetric rain gauge. The rain is accumulated in a collection funnel equipped with a self-initiating siphon that empties the gauge in 20 s when the funnel becomes full. The water level inside the funnel is sensed with a capacitance-type, water-level transducer with a resolution of 1 mm. The water level is sampled every 6 s, and 60-s averages are recorded. Rain rate is derived by first differencing the water level in time. The large negative changes in water level associated with the siphoning of the gauge are removed.

e. Sea surface temperature

Sea surface temperature was measured by both an IMET sea surface temperature module deployed at 1.0 m and a Brancker Model XL100 self-contained temperature logger deployed at 0.45 m below the surface. The Brancker T-pod temperatures are used here for sea surface temperatures since they are located closer to the surface. This logger was the shallowest of six temperature loggers deployed in a near-surface temperature array attached to the buoy hull (Anderson et al. 1996). Sea surface temperature sensors calibrate in the laboratory to an accuracy of better than 0.01°C. In the field, intercomparisons have demonstrated a sea surface temperature accuracy of approximately 0.1°C (Weller and Anderson 1996).

3. Observations

As part of TOGA COARE, the PTM was deployed in the western Pacific warm pool on the WHOI surface mooring. The mooring was deployed at the center of the warm pool (1°45′S, 156°E) from 21 October 1992 to 4 March 1993. The 3-m discus buoy carried a complete IMET package that sampled wind velocity, relative humidity, air temperature, barometric pressure, incoming shortwave radiation, incoming longwave radiation, sea surface temperature, precipitation, and, with the PTM, precipitation temperature every minute (Weller and Anderson 1996). The initial deployment occurred on 21 October 1992, when the PTM operated for 14 days. On 13 December 1992, the surface buoy was outfitted with a second suite of IMET sensors, and the PTM operated for five days. The logger failed prematurely on both deployments from an electrical power surge that likely was caused by lightning.

The observations include six significant precipitation events, each with cumulative freshwater fluxes of more than 5 mm occurring within 2 h (Fig. 2). These occur on 21 October, 27 October, 28 October, 13 December, 15 December, and 16 December 1992 (see Table 1). There are also several small events in late October and early November. The events that occurred in October correspond to a period of small-scale atmospheric convection in the region. The events in December correspond to the onset of large-scale atmospheric convection and a series of westerly wind bursts (Weller and Anderson 1996).

Two of the events are examined closely to illustrate the temporal evolution of the precipitation temperatures in relation to the wet- and dry-bulb temperatures, the sea surface temperature, and the rain rate. Event 3 actually consists of two closely spaced events (Fig. 3). There is an initial drop in the air temperature, Tair, from 28° to 26°C, 10 min prior to the onset of rain. Then Tair rapidly drops to a minimum of 24°C following the start of the rain. The dry PTM temperatures have a spread of 0.5°C that are grouped just below ambient air temperature before the event. The PTM temperatures then become closer together when the rain starts and drop down below Tair and Twet at 23.3°C. Shortly after, all the PTM temperatures begin to rise to Twet. After this first burst of rain, the PTM temperatures continue to rise and spread apart. The warmest is PT1, which is suspended in the center of the funnel and has little thermal mass and thus can adjust more quickly back to ambient air temperature than the rest of the unit. The onset of the second burst of rain brings Tair and PTM temperatures down to Twet. They all group together within 0.25°C for the duration of the rainfall. Note that Twet increased during the first event and decreased during the second, and this variability is matched by the PTM temperatures. There is a 0.1°C drop in the sea surface temperature during the event.

The start of event 4 is very similar to event 3, with the air temperature dropping rapidly prior to the onset of rain (Fig. 4). The PTM temperatures and Tair drop down to Twet at the onset of rain and then remain closely matched to Twet for the duration of the event. The Twet temperature decreases during the length of the event by approximately 1.0°C. The sea surface temperature falls 0.3°C during the event and is on average 5.2°C warmer than the PTM temperatures.

The October events occur during a period of small-scale convective events in the COARE region, while the December events occur during the onset of large-scale atmospheric convection (Weller and Anderson 1996). The December events are associated with high relative humidities, while the October events have significantly lower relative humidities. The event-averaged temperatures for each of the six events are given in Table 1. The precipitation temperature (PT) is equal to Twet in event 2 and 0.1°C lower than Twet for all of the other events. Event 2 clearly is different from the other events. It has the lowest relative humidity and wind speed and the highest instantaneous rain rate, lasting for less than 25 min, and is the only event to occur during daylight hours. The other events have durations of more than an hour and have average winds ranging from 4.6 to9.9 m s−1. The rain rates range from 10.8 to 31.3 mm h−1. The sea surface temperature for the events range from 29.0° to 29.9°C and are 4.8°–5.8°C above the PTM temperatures.

The observations yield droplet temperatures that are on average slighly lower than the wet-bulb temperature. The result that the droplet temperatures are cooler than Twet supports the hypothesis of Kinzer and Gunn (1951), Kincaid and Longly (1989), and Gosnell et al. (1995). However, with the large uncertainty in the wet-bulb temperature measurement, this result is not significant. The accuracy in the rain droplet temperature measurement is higher than the accuracy of the wet-bulb temperature estimate (±0.4°C). The fact that they are so close tells more about the accuracy of the humidity measurements than about the rain temperature. The result is the equivalent of an in situ humidity calibration that increases our confidence in the relative humidity measurement that is likely better than ±3% accuracy.

Improvements in the long-term stability and accuracy of the precipitation temperature measurements could be enhanced by using platinum resistance thermometers (PRTs) in conjunction with fast-response thermistors. PRTs are inherently more stable for longer periods of time than thermistors, but thermistors have a much faster response time. By using both in the measurement, the long-term stability can be improved along with maintaining the quick response. The electronics could be redesigned to make the PTM completely self-contained by reducing the power consumption, internally self-recording, and using an internal battery pack. A different choice of materials, such as aluminum, for the funnel could also yield improvements in the response time of the PTM to changes in precipitation temperature.

4. Discussion

What is the relative significance of including the sensible heat flux due to rain in net heat and buoyancy flux calculations?

The four components typically included in the calculation of the surface heat flux are the net shortwave and longwave radiative fluxes and the latent and sensible heat fluxes. A complete description of the observed fluxes and their derivation from the moored observations during COARE is reported by Weller and Anderson (1996). The empirically derived transfer coefficients and boundary layer profiles used in the bulk flux formula are those of Fairall et al. (1996). Due to a lack of observational data, it is not known whether these coefficients adequately represent surface roughness and the boundary layer profiles of wind speed, specific humidity, and temperature while it is raining. Therefore, there may be significant errors when using these bulk aerodynamic formulas to calculate the sensible and latent heat flux while it is raining. In addition, the rain likely blocks the longwave radiometric measurement by coating the dome of the sensor with water (which is opaque to longwave radiation), and the sea surface longwave emissivity is not known while it is raining. These limitations on heat flux estimates are overcome since typically the portion of time that it is raining is small and the errors will be minor when the fluxes are averaged over periods much longer than the individual rain events. However, there are clearly uncertainties in measuring the surface heat flux when it is raining.

The sensible heat flux associated with rain, HR, is calculated following the formulation of Flament and Sawyer (1995)
HRCprρTrainTsr,
where Cpr is the specific heat capacity of freshwater, ρ is the density of freshwater, Ts is the sea surface temperature, and r is the rain rate. The uncertainty in HR comes mainly from the uncertainty in the rain rate and rain or wet-bulb temperature measurements. An error in the wet-bulb temperature of 0.5°C corresponds to approximately a 10% error in HR for these observations. The mean relative error in rainfall rate from the capacitance rain gauge is inversely proportional to rainfall rate. For 1-min average rainfall rates above 5 mm h−1, the error is 14% (Nystuen et al. 1996). Thus, the HR error is approximately 25% and may have the smallest uncertainty of all the components of the heat flux while it is raining. The HR as well as the other four components of the surface heat flux for each of the six events are given in Table 2. The rain heat flux accounts for 15%–60% of the net heat flux while it is raining and is clearly an important component to the surface forcing during a rain event.

In the western tropical Pacific, the rain heat flux has an impact on longer timescales as well (Gosnell et al. 1995). The net surface heat flux, excluding the rain heat flux, during the 4-month mooring deployment is 17.5 W m−2 (ocean warming) in the warm pool region (Weller and Anderson 1996). The mean rain rate for these four months is 10.3 mm day−1. Assuming a mean rain droplet–sea surface temperature difference of −5°C, the mean rain heat flux, calculated from (5), is 2.6 W m−2, or 15% of the net surface heat flux. Even though on average the rain heat flux is the smallest term in the surface heat budget, it should be included to fully resolve the close balance between surface heating and cooling.

The buoyancy forcing on the surface mixed layer in the warm pool plays an important role in setting the upper-ocean characteristics in the region (Anderson et al. 1996). The surface freshwater flux from rain decreases the salinity of the ocean surface, thus increasing the hydrostatic stability. The sensible heat flux of rain works against the freshwater flux by cooling the surface waters and reducing hydrostatic stability. What is the net effect on the buoyancy flux from rain? The buoyancy fluxes for each of the six events are given in Table 3. The net buoyancy flux from rain is written as
i1520-0426-15-4-979-e6
where α and β are the thermal and haline coefficients of expansion, and S0 is the reference surface salinity. Assuming that the rain droplet–sea surface temperature difference is −5°C, the reduction of the buoyancy flux due to rain by the cooling effect of rain is 6.5%. As expected, the buoyancy flux from the rain dominates all the buoyancy flux terms while it is raining. However, the reduction in the buoyancy forcing due to the sensible heat flux from rain is an important term when the net buoyancy flux is small, as in events 1 and 6.

5. Conclusions

A new instrument has been developed to make direct observations of precipitation temperature from a surface buoy. The instrument was deployed in the western equatorial Pacific warm pool during TOGA COARE. Previous work indicates that the rain droplet temperature will be near or just below the surface wet-bulb temperature. The observed droplet temperatures are equal to the wet-bulb temperature to within the uncertainty in wet-bulb temperature measurement (±0.4°C). The sensible heat flux associated with the rain is found to be a significant component to the net surface heat flux in the warm pool region. It accounts for up to −204 W m−2 during a single rain event and for −2.6 W m−2 over the full 4-month deployment. The surface buoyancy flux from the rain is reduced by 6.5% when the effects of the sensible heat flux of rain are included.

Acknowledgments

Our participation in TOGA COARE was supported by the National Science Foundation, Grant OCE91-10559. Assistance from the members of the Upper Ocean Processes Group during all phases of the work is gratefully acknowledged. Mary Ann Lucas assisted in the preparation of this manuscript.

REFERENCES

  • Anderson, S. P., and M. F. Baumgartner, 1997: Radiative heating errors in naturally ventilated air temperature measurements made from buoys. J. Atmos. Oceanic Technol.,15, 157–173.

    • Crossref
    • Export Citation
  • ——, R. A. Weller, and R. B. Lukas, 1996: Surface buoyancy forcing and the mixed layer of the western Pacific warm pool: Observations and 1D model results. J. Climate,9, 3056–3085.

    • Crossref
    • Export Citation
  • Buck, A. L., 1981: New equations for computing vapor pressure and enhancement factor. J. Appl. Meteor.,20, 1527–1532.

    • Crossref
    • Export Citation
  • Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S. Young, 1996: Bulk parameterization of air–sea fluxes for TOGA COARE. J. Geophys. Res.,101 (C2), 3747–3764.

    • Crossref
    • Export Citation
  • Flament, P., and M. Sawyer, 1995: Observations of the effect of rain temperature on the surface heat flux in the intertropical convergence zone. J. Phys. Oceanogr.,25, 413–419.

    • Crossref
    • Export Citation
  • Gill, G. C., 1983: Comparison testing of selected naturally ventilated solar radiation shields. NOAA Data Buoy Office Report, Contract NA-82-0A-A-266, NOAA/National Data Buoy Center, Bay St. Louis, MS, 15 pp.

  • Gosnell, R., C. W. Fairall, and P. J. Webster, 1995: The sensible heat of rainfall in the tropical ocean. J. Geophys. Res.,100 (C9), 18 437–18 442.

    • Crossref
    • Export Citation
  • Hosom, D. S., R. A. Weller, R. E. Payne, and K. E. Prada, 1995: The IMET (Improved Meteorology) ship and buoy system. J. Atmos. Oceanic Technol.,12, 527–540.

    • Crossref
    • Export Citation
  • Kincaid, D. C., and T. S. Longley, 1989: A water droplet evaporation and temperature model. Trans. ASAE,32, 457–463.

    • Crossref
    • Export Citation
  • Kinzer, B. D., and R. Gunn, 1951: The evaporation, temperature, and thermal relaxation-time of freely falling waterdrops. J. Meteor.,8, 71–83.

  • List, R. J., 1984: Smithsonian Meteorological Tables. Smithsonian Institution Press, 527 pp.

  • Nystuen, J. A., J. R. Proni, P. G. Black, and J. C. Wilkerson, 1996: A comparison of automatic rain gauges. J. Atmos. Oceanic Technol.,13, 62–73.

    • Crossref
    • Export Citation
  • Webster, P. J., and R. Lukas, 1992: TOGA COARE: The Coupled Ocean–Atmosphere Response Experiment. Bull. Amer. Meteor. Soc.,73, 1377–1416.

    • Crossref
    • Export Citation
  • Weller, R. A., and S. P. Anderson, 1996: Surface meteorology and air–sea fluxes in the western equatorial Pacific warm pool during TOGA Coupled Ocean–Atmosphere Response Experiment. J. Climate,9, 1959–1990.

    • Crossref
    • Export Citation

Fig. 1.
Fig. 1.

Diagram of the precipitation temperature module. It indicates the location of the four thermistors: PT1, PT2, PT3, and PT4.

Citation: Journal of Atmospheric and Oceanic Technology 15, 4; 10.1175/1520-0426(1998)015<0979:MOOPT>2.0.CO;2

Fig. 2.
Fig. 2.

Observed precipitation. The observed rain rate (black line) and the cumulative rainfall (gray line) are shown for the two deployments. The data represent 1-min averages.

Citation: Journal of Atmospheric and Oceanic Technology 15, 4; 10.1175/1520-0426(1998)015<0979:MOOPT>2.0.CO;2

Fig. 3.
Fig. 3.

Rain rate and droplet temperatures for event 3. Data shown are for 5.5 h during 28 October 1992. (a) The observed rain rate (black line) and the cumulative rainfall (gray line) are shown for the two deployments. (b) The four temperatures from the PTM are plotted for the corresponding time period (thin black lines). Also shown are the sea surface (dashed), dry-bulb (thick black), and wet-bulb (thick gray) temperatures. The data represent 1-min averages.

Citation: Journal of Atmospheric and Oceanic Technology 15, 4; 10.1175/1520-0426(1998)015<0979:MOOPT>2.0.CO;2

Fig. 4.
Fig. 4.

Rain rate and droplet temperatures for event 4. Same as Fig. 3 except data are from 5 h on 13 December 1992.

Citation: Journal of Atmospheric and Oceanic Technology 15, 4; 10.1175/1520-0426(1998)015<0979:MOOPT>2.0.CO;2

Table 1.

Event statistics. The data represent means over each of the six 1992 events discussed in the text.

Table 1.
Table 2.

The heat flux components for each event in 1992 (W m−2).

Table 2.
Table 3.

Components of buoyancy flux for each event [(kg s−1 m−2) × 10− 6].

Table 3.

* Contribution Number 9280, Woods Hole Oceanographic Institution.

Save