A Simulation Study of the Bias Error Analysis of Mean Rainfall Rates Measured with Spaceborne Radar

Abstract

Spaceborne rain radars often operate with a fairly low SNR (signal-to-noise ratio) because of several hardware limitations and the large distance between the radar and target. Rain radar do not directly measure the rain echo power but instead the total received power, which is the sum of the rain echo power and the receiver noise power. The rain echo power is obtained by taking the difference between the total received power and the receiver noise power. The experimentally obtained rain echo power is sometimes a negative value when the SNR is low. The rainfall rate cannot be directly estimated from this negative data. This paper proposes three data processings (truncation, zero rain, and negative rain methods) for mean rainfall rate estimations from rain echo data that includes data with a negative value. The bias error analysis for the three methods is done by computer simulation. The bias error from truncation is fairly large compared with that caused by the other two methods. When the receiver noise power is small, the zero rain method, which gives relatively small bias error, will be useful. When the receiver noise power is large, the negative rain method, which effectively reduces the bias error, will be useful. Another feature of the negative rain method is that it automatically cancels the effect of misidentification of rain or no-rain.