MODAPS PCR 17-029
| PCR Number | 17-029 |
|---|---|
| Date | 2017-12-21 |
| Initiator | Hulley, Glynn C. |
| Abstract | PGE116 v6.0.63 -- - Fix the code reading the LUNs for processing parameters - Change code to measure height in km - Initialize variable “count” to fix error estimate - Change cloud processing to choose c95 instead of c66 - Set QC bits 1 and 0 to 10 when LST > 400K - Use different PWV threshold and smoothing scale when Tgmean < 295K - Clamp gi to (-2.0 3.0) - Change ASTER minimum from 0.7 to 0.8 - Fix code to correct metadata |
| Problem Statement | Data from the operational C6 processing had many flaws: - Water vapor model mostly turned off resulting in large positive bias in the retrieved LST - Inaccurate temperature and emissivity at higher elevations - LST error estimates were wrong for a large number of retrievals - LST retrieved for 66% cloud clear state and were not reliable |
| Description of Change | Fixed software bugs to address the above issues: - Fix the LUN identifying the correct variable to turn on the water vapor scaling model - Code change to identify and use height data in km (not meters) as expected by the RTTOV to fix the inaccuracies in the temperature and emissivity at higher elevations - Variables properly initialized to fix the LST error estimates - LST retrieved for 95% cloud clear state and better Additional minor code changes included to improve the accuracy of the retrieved LST |
| Products Affected | Daily L2 LST – MO(Y)D21 |
| Software Affected | PGE116 v6.0.63 |
| Processing String to Receive the Change | Terra and Aqua Forward Processing. Terra and Aqua Reprocessing. NRT |
| Downstream Product Effects | None. |
| Data Granules to be Used for Science Testing | Run test over selected days in 2006 where the issue was first seen and verified by LDOPE. |
| Justification | Validation of the product (MxD21 suite) from the operational C6 processing revealed large bias and inaccuracies in the retrieved LST during moist and humid conditions. Examination of the code revealed that a LUN identifying the threshold used to switch on the water vapor scaling model was incorrect. Accuracy of LST retrieved using the Terra MODIS observation was also negatively impacted by the cross talk in Terra MODIS B29. Without the re-reprocessing of the product the C6 LST product (MxD21) series would be largely unusable during conditions with large water vapor loading. Science team recommends reprocessing of the Terra MODIS for the period 2000 -2008 and the full mission of Aqua. It is recommended that LST for the Terra period 2008 and beyond be not generated in C6 as the product will not be science quality due to the B29 crosstalk problem. |
| Effective Date for Implementation of Change | Completed. |
| Status | Major algorithm development completed. Sadashiva Devadiga approved on 12/21/2017. Michael King approved 12/29/2017. Gang Ye implemented 01/02/2018. |
