Major changes in observing condition

Name of standard product after the boost

1B21 Algorithm

Changes in 1B21 algorithm
1B21V5A Warning on mainlobe clutter rejection

1C21 Algorithm

2A21 Algorithm

2A23 Algorithm

2A25 Algorithm

3A26 Algorithm

Major changes in observing condition 　N. Takahashi

　Sensitivity of the Precipitation Radar (PR) to measure radar echoes near Earth's surface is lower by about 1.2dB after the boost than before because of the increase in distance to the surface. The proportion of weak rain echoes that cannot be detected by PR has increased. 　The footprint diameter has increased by 15% and its area became 1.3 times the area before the boost. The across-track swath width became about 245 kmfrom the nominal swath width of 215 km at 350-km altitude.

　The highest vertical sample altitude was lowered by 1.5 km at nadir.　Beam mismatch between the transmitter and receiver antenna angles occurs for two pulses among 64 averaging pulses. The data obtained under the mismatched condition is equivalent to the data obtained with the antenna whose direction for both transmission and reception is half-way between the mismatched transmitting and receiving directions and with 6dB gain reduction. PR transmits 64 pulses for each scan angle, and averages the corresponding echoes with an onboard processor. 　The incident angle changed slightly, especially for large scan angles. These changes cause minor degradation of data quality and apparent changes in some statistical values (e.g., sigma zero statistics). "Beam mismatch" increases uncertainty in the Z values.

Name of standard product after the boost 　N.koutaki

　In order to distinguish from the pre-boost 350km data, the post-boost products are marked with "5A" as the product version number in the meta data. This new product designation is used after orbit number 21259 (inclusive) data of August 7, 2001.

Changes in 1B21 algorithm N. Takahashi

The new algorithm (5A) is written only for 402.5-km altitude data.

1. A new table on the first sampling range bin distance is set for the 402.5-km algorithm.

2. Alternative system noise sampling is provided for some angle bins in case a very tall echo contaminates the nominal system noise sample.

3. A beam mismatch correction algorithm is added. This algorithm assumes that the mismatched pulsed data after some gain adjustment is represented by the average power of the echoes received independently at the two mismatched beam directions under the matched beam condition.

   The maximum estimated error from this algorithm is about 0.7 dB, which may appear in the surface echo, but the typical error of rain echo is less than 0.1dB.

4. The search window of the main lobe clutter routine was changed slightly because the incident angles changed slightly for the 402.5-km operation. This may cause slight changes in "binClutterFreeBottom" in 1B21.

1B21 V5A Warning on mainlobe clutter rejection　J. Awaka

　Rejection of mainlobe clutter in 1B21 V5A works as well as in V5. This means, however, that in certain circumstances, e.g., over high mountain areas, a strong mainlobe clutter can be misjudged as a strong rain echo, though the frequency of this misjudgment is very low.

1C21 Algorithm　S. Shimizu

　No changes were made in the 1C21 code with the boost. The minimum detectable value of Z-factor after the boost is 1.2dB higher than before because the sensitivity of PR has decreased by 1.2 dB.

2A21 Algorithm R. Meneghini

　No changes were made in the algorithm.
　Because of the effects of the earth's curvature and the increased swath, the non-nadir incidence angles with respect to the surface normal increase slightly.　After the orbit boost, the attitude of the spacecraft appears to be less reliable so that the　estimate of the incidence angle with respect to the local normal is also less reliable.

　This uncertainty translates into an error in the classification of the NRCS (normalized radar cross section) of the surface and a larger variance in the non-raining NRCS data. The increase in altitude causes a loss of synchronization between transmission and reception of the last pulse pair (1 of 32 pulse pairs). Since the mean values of the NRCS
have changed slightly from the pre-boost means, and that this change is a function of the surface type and incidence angle, the correction for this effect may be in error.

　Because of a loss in sensitivity of about 1.2 dB, the rain/no-rain threshold has changed. This will slightly modify the rain-free statistics of the NCRS and will also modify the mean path-integrated attenuation as estimated by the surface reference technique.

　Over ocean at near-nadir incidence the rain-free NRCS post-boost values are slightly higher than pre-boost values. Near the swath edge the post-boost NRCS values are lower by 0.3-0.5 dB.

2A23 Algorithm J. Awaka

　The success rate of bright band (BB) detection decreases with increased antenna scan angle because of the smearing of the BB profile. This tendency becomes stronger in V5A. When 2A23 fails to detect BB, a large BB peak is sometimes misinterpreted as a strong convective echo. Chances of misjudging BB as a strong convective echo increase in V5A in particular for the data near the edges of the swath.

2A25 Algorithm　T. Iguchi

　The algorithm of 2A25 after the boost is exactly the same as that before the boost. No code has been changed. (All known bugs remain in the code.) However, there are
some changes in the quality of data after the boost.

1. The altitudes of the lowest valid data near the scan edges have increased in accordance with the increase of the footprint size. Now, the height of 2 km is not high enough to be free from surface clutter over the ocean.

2. Although the non-uniform beam filling effect must have increased after the boost, no modification or adjustment of coefficients was made to accommodate this change.

3. Since the threshold of the detectable signal has increased by about 1.2 dB, the PR misses more weak rain cases. This change will increase the mean rainfall rate conditioned on rain. Be careful when taking statistics. Also, some noise that exceeds the threshold (and hence is misidentified as rain echo) is now converted to a largerrain rate than before because of the change in the conversion factor from the in the conversion factor from the received power to the radar reflectivity factor. The proportion of this kind of misidentification remains approximately the same. Because such misidentified noise signals are relatively predominant at high altitude, the statistical means of rain rate conditioned on rain at high altitudes appear 　conditioned on rain at high altitudes appear to have increased after the boost, but they are not real.

4. Look at the caveats for other level 1 and level 2 PR products.

3A25 Algorithm 　　R. Meneghini

　No changes were made in the 3a25 code. However, since 3a25 calculates the statistics of the level 1 and 2 data the outputs of 3a25 will reflect changes in the lower level products. In particular, we expect the conditional mean rain rates and radar reflectivity factors to increase slightly and the unconditional particularly for low-level rains. Some rain data that formerly would have been detected at a height of 2 km will now be obscured by surface clutter.

　The input data set for the 3a25 August 2001 product consists of the level 1 and 2 data sets taken before (1-7 August) and after (25-31 August) the boost. Because the product is based on only fifteen days of data, the sampling errors in the products will be larger than typical values.

3A26 Algorithm　　R. Meneghini

　3a26 computes the fractional area coverage of rain at various rain rate thresholds and estimates the space-time averaged rain rate based on a log-normal model. Because of the decrease in sensitivity, the fractional area coverage at low rain rates will decrease. The space-time averaged rain rate is expected to change because of the dependence of the log-normal fitting procedure on the distribution at low rain rates. No changes were made to the 3a26 code.