Effects of Rainfall on WatER
Ernesto RodriguezJet Propulsion Laboratory
California Institute of Technology
Effects of Rainfall on WatER
• Water is a Ka-band instrument• Ka-band radiation is strongly attenuated in the
presence of rain• Unlike nadir altimeters, the main effect of rain on the
radiometer will be a loss of signal-to-noise ratio, withan increase in height noise.
• Currently, the WatER budget allows for meeting thescience requirements with a drop in power of 6dB
• There is also a concern about rainfall data loss beingcorrelated with time of day if a sun-synchronous orbitis used
Attenuation Effects at Ka-Band
6 dB311.5 dB5
215 dB10073 dB3050 dB2024 dB10
35 GHz Attenuation(5-km path)
Rain rate mm/hr
WatER will only be able to collect valid data at rainrates smaller than 3-5 mm/hr (depending onsurface water brightness)
Walsh, et al., Rain and Cloud Effects on aSatellite Dual-Frequency Radar AltimeterSystem Operating at 13.5 and 35 GHz,IEEE Trans. GRS, 22, 1984
Rain Probability
Petersen WA, Nesbitt SW, Blakeslee RJ, et al.TRMM observations of intraseasonal variability in convective regimes over the Amazon JOURNAL OF CLIMATE 15 (11): 1278-1294 JUN 1 2002
The total WatER data loss, if data were uncorrelated with time of day,will be less than 10% in the tropics. Similar numbers also hold at otherlatitudes.
Data Loss vs Time of Day
• In non-tropical regions, the correlation between time of day andrain events is weak– Rainfall dominated by fronts– Data loss expected to be < 10%
• Tropical rainfall is governed by convective instabilities whicharise due to daytime heating– Thunderstorms tend to happen in the late afternoon or
evening– Rainfall pattern is patchy– Correlation distance between rain events < 50 km to 100 km– 6 am is near the rainfall probability minimum, while 6 pm is
close to the rainfall probability maximum– See results below for more details
Spatial Distribution of Rain Cells-1
Spatial Distribution of Rain Cells-2
Monthly Weather Review, 134, 2004
Note localized rain cells with patchydistribution
Seasonal Variation of Precipitation
Monthly Weather Review, 129, 2001
Local Time of Maximum Precipitation
Monthly Weather Review, 129, 2001
Evolution of Rain Rate vs Local Time-NW South America-1
Monthly Weather Review, 131, 2003
Evolution of Rain Rate vs Local Time-NW South America-Rain Gauges
Monthly Weather Review, 133, 2005
Rainfall vs Time of Day
Journal of Climate, 16, 2003
Amazon/LBASoutheast US
Monthly Weather Review, 108, 1980
Conclusions
• There will likely be a global data loss <10% due rainfall• The data loss away from the tropics is not correlated with time
of day in a significant way• In the tropics, the likelihood of rain for the 6pm passes is
significantly higher than for the 6am passes. However,averaging between the two, a 10% data loss is still expected.
• Temporal sampling in the tropical regions may be moreimpacted than away from the tropics due to greater loss of datain the 6pm passes
• However, the temporal sampling will still meet the samplingrequirements for the mission
• A 10% data loss is acceptable for spaceborne remote sensingmissions: e.g., Topex or Jason have 81% data recoveryrequirements
• Due to power considerations, a 6am/6pm sun-synchronousorbit is still the most attractive possibility.