S-band calibration offsets as determined for each period of stable operation (NOT APPLIED TO LEVEL 0 OR LEVEL 2 DATA - users need to apply these offsets if needed)

Start Date   Stop Date   Calibration Offset (dB)
1/3/2014     2/5/2014    + 1.0
2/6/2014     8/19/2014   + 3.0
8/20/2014    10/16/2014  - 2.5
10/17/2014   3/6/2015    - 5.5
3/7/2015     7/5/2015    - 4.0
7/6/2015     10/28/2015  - 1.5
10/29/2015   12/31/2015  + 1.0

The calibration constants were based on a comparison to TRMM PR data (Schumacher and Houze 2000). A TRMM comparison to the longest period of stable radar operations (10/17/2015 - 3/6/2015) provided the baseline value for calibration offsets and then other offsets were determined by comparison to this baseline. Stable periods of radar operation were determined through a study of the 95th percentile values of ground clutter pixels (Silberstein et al. 2008).

CAPPI procedure

CAPPIs were derived using the NASA-TRMM RSL toolkit available at https://github.com/tsbiscaro/rsl-v1.50_HDF5

Level 2 data are corrected reflectivity (i.e.: w/ clutter filter) CAPPIs from 2 up to 16 km, one CAPPI each 1 km. The data are 4 bytes (floating point), 500 x 500 x 15, i.e., 15 blocks (one for each level) of 500 x 500 grid points. The first matrix is the 2 km CAPPI, then the 3 km, up to the 16 km CAPPI.

Undefined values are signed with -99

The horizontal resolution is 1 km, 250 km radius.

Known Issues
Number of sweeps can vary from volume to volume
The volume scan created at SIPAM Manaus should include 17 sweeps per volume, but due to internet connectivity issues and other, the number of sweeps in a volume is variable across the data set. When sweeps are missing from a volume, they are missing in a top-down fashion. For example, if sweeps are numbered 1 to 17, a volume could be missing sweep 17, sweeps 16 to 17, sweeps 15 to 17 and so forth. The variable number of sweeps from volume to volume has implications for the CAPPI dataset. The area of missing data near the center of the CAPPI grid can be larger due to missing sweeps. Further, echo top-height results will be affected by missing sweeps.


Ground clutter due to anomalous propagation
The data can also have large areas of ground clutter around the radar site possibly due to anomalous propagation of the radar beam in certain scenarios. The success of the updated quality control method in identifying the AP echo varies across the dataset. Some periods of radar data still suffer from significant AP echo, while AP echo has been almost completely censored during other periods. 


Beam blockage 
It is evident in the long-term rain statistics that some radar beams are at least partially blocked.


November 2014 missing data
During early November of 2014 (Nov. 3 to Nov. 11), several days of noisy data exists.


Sweeps with Reduced Number of Rays
The number of rays per sweep can be less than 360. This change reduces the sampling area of the radar volume and reduces the number of grid points with data in the interpolated fields. 


Antenna seeking during scanning
The SIPAM radar antenna elevation transitions are recorded in the data so that sweeps contain rays with elevation angles with significant error when compared to the sweep’s target angle