IR Based Precipitation Estimates IR Based Precipitation Estimates as an Input to CMORPH as an Input to CMORPH

(1)

IR Based Precipitation Estimates IR Based Precipitation Estimates

as an Input to CMORPH as an Input to CMORPH

Pingping Xie Pingping Xie

NOAA Climate Prediction Center NOAA Climate Prediction Center

2018.11.23.

(2)

Objective Objective

• To introduce CPC Precipitation Team’s efforts to develop satellite IR based precipitation estimates as inputs to the second generation CMORPH

integrated satellite pole-to-pole global

precipitation estimates;

(3)

Why We Need IR Precipitation Estimates Why We Need IR Precipitation Estimates

CMORPH2 Flowchart CMORPH2 Flowchart

• Geostationary IR based

precipitation estimates in 30-min interval are used to derive cloud motion vectors

• Geostationary IR based

precipitation estimates are used to fill in the gaps of passive

microwave (PMW) retrievals, especially in real-time production before PMW retrievals are

available from sufficient number of sensors

• Low earth orbit (LEO) IR based precipitation estimates are used to fill in gaps over the mid- and high latitudes where current generation technology is unable to derive

retrievals of precipitation from PMW measurements;

(4)

Why We Do Not Use Why We Do Not Use

IR Technique Developed by Other Centers IR Technique Developed by Other Centers

• Geostationary IR based precipitation estimates

• The IR based precipitation estimates need to present close PDF agreement with the inter-calibrated PMW retrievals in the

CMORPH system

• The grid system needs to fit into our own

• Real-time production requirements \

• We always want to get the best possible products

• LEO IR Precipitation Estimates

• Our emphasis is a satellite precipitation product with reasonable quality during cold season and over cold surface to fill in the gaps of PMW retrievals;

• Performance for tropical and warm season precipitation is secondary;

• No LEO IR precipitation products have been targeted for this specific application

(5)

GEO IR-Based Precipitation Estimates [1]

The original IRFREQ technique The original IRFREQ technique

 Developed over 10 years ago [Joyce et al 2004]

 TBB – precip relationship defined through PDF matching against MWCOMB

 Considering land / sea differences and latitudinal changes in the TBB – precip relationship (no zonal variations considered)

 Data over a 9-hour window centering at the target

analysis time used to account temporal variations of the relationship (no data over recent days included)

(6)

GEO IR-Based Precipitation Estimates [2]

Modification to the IRFREQ for post-process Modification to the IRFREQ for post-process

 Still based on PDF matching against MWCOMB

 TBB – precip relationship established for each

1^olat/lon grid using data over a 7-hour window centering at the target hour and for a 15-day period ending at the target date

 Performed for both GEO and LEO IR data

1-5 March, 2014 1-5 March, 2014

MWCOMB MWCOMB

LEO-IR-PRCP LEO-IR-PRCP

(7)

GEO IR-Based Precipitation Estimates [3]

Modification to the IRFREQ for real-time processing Modification to the IRFREQ for real-time processing

 Problem:

PDF matching using data up to a day ago can not capture the rapidly evolving TBB – precip

relationship sometime.

 Bottleneck:

MWCOMB is available several hours later and therefore can not be used to update the TBB –

precip relationship in real-time processing

 Solution:

Using QMORPH to replace

MWCOMB for recent hours for which MWCOMB is not available

14Z, 18 July, 2014

(8)

GEO IR-Based Precipitation Estimates [4]

Real-time GEO-IR based precip estimates Real-time GEO-IR based precip estimates

(9)

Percentage of hours Percentage of hours during DJF of 2017-2018 during DJF of 2017-2018 covered by (left) PMW covered by (left) PMW retrievals

retrievals of rainfall rate of rainfall rate (RR); (middle

(RR); (middle) PMW ) PMW retrievals of rainfall rate retrievals of rainfall rate (RR) and snowfall rate (RR) and snowfall rate (SFR), and (right) all (SFR), and (right) all precipitation retrievals precipitation retrievals including PMW based including PMW based RR, SFR, as well as LEO RR, SFR, as well as LEO IR (AVHRR) based

IR (AVHRR) based estimates.

estimates.

Q:Q: Why do we we still need LEO IR based precipitation estimates when we already have PMW Why do we we still need LEO IR based precipitation estimates when we already have PMW retrievals?

retrievals?

A:A: Because there are still a lot of missing gaps in the current generation PMW retrievals over Because there are still a lot of missing gaps in the current generation PMW retrievals over the high latitudes. We need something (much) better than nothing to fill in the gaps

the high latitudes. We need something (much) better than nothing to fill in the gaps

PMW RR PMW RR+SFR PMWRR+SFR+AVHRR P PMW RR PMW RR+SFR PMWRR+SFR+AVHRR P

LEO IR-Based Precipitation Estimates [1]

Why We Need LEO IR Precip Estimates Why We Need LEO IR Precip Estimates

(10)

25-29 Jul 2017 30 Jul – 3 Aug 2017 4–8 Aug 2017 9-13 Aug 201725-29 Jul 2017 30 Jul – 3 Aug 2017 4–8 Aug 2017 9-13 Aug 2017

• AVHRR TB corrected for limb effects (figures skipped)AVHRR TB corrected for limb effects (figures skipped)

• PDF matching of AVHRR against MWCOMB and PDF matching of AVHRR against MWCOMB and CloudSat radar estimates

CloudSat radar estimates

• Calibration tables established as a function of region Calibration tables established as a function of region (1(1^o^olat/lon grid), season (pentad), and surface type;lat/lon grid), season (pentad), and surface type;

• Fine-tuning for optimal performance over polar regions;Fine-tuning for optimal performance over polar regions;

LEO IR-Based Precipitation Estimates [2]

Why We Did It Why We Did It

(11)

MWCOMB PMW RR (mostly NESDIS) MWCOMB PMW RR (mostly NESDIS)

1 - 10 Dec 2017 [accum mm]

MWCOMB GPROF-V5 surface precipitation MWCOMB GPROF-V5 surface precipitation

1 – 10 Dec 2017 [accum mm]1 – 10 Dec 2017 [accum mm]

APCOMB (PMW RR+SFR+AVHRR) APCOMB (PMW RR+SFR+AVHRR)

1 - 10 Dec 2017 [accum mm]

• Inter-calibrating inputs from various Inter-calibrating inputs from various sources against a common

sources against a common

reference standard through PDF reference standard through PDF matching

matching

• Constructing composite fields of Constructing composite fields of inter-calibrated precipitation fields, inter-calibrated precipitation fields, called

called APCOMBAPCOMB, on a 0.05, on a 0.05^o^olat/lon lat/lon grid and in 30-min intervals

grid and in 30-min intervals

LEO IR-Based Precipitation Estimates [3]

What We Have Achieved What We Have Achieved

(12)

• Done several years ago as part of our GOES-R Done several years ago as part of our GOES-R Risk Reduction Program Project:

Risk Reduction Program Project:

• Develop GOES-R enhanced regional CMOROHDevelop GOES-R enhanced regional CMOROH

• Further combine the regional CMORPH with radar and gaugeFurther combine the regional CMORPH with radar and gauge

• GOES-R enhanced CMORPHGOES-R enhanced CMORPH

• 2kmx2km / 10-min resolution 2kmx2km / 10-min resolution

• reduced latency reduced latency

• improved quantitative accuracy improved quantitative accuracy

• Key to the GOES-R enhanced CMORPHKey to the GOES-R enhanced CMORPH

• GOES-R based precipitation estimates (GPE)GOES-R based precipitation estimates (GPE)

• derived from IR (band-13) and other channels including lightning derived from IR (band-13) and other channels including lightning information

information

Preliminary Work on H8 IR Precip Estimates Preliminary Work on H8 IR Precip Estimates

[1] Background [1] Background

(13)

• We have developed an algorithm to derive We have developed an algorithm to derive precipitation estimates from the GOES-R IR precipitation estimates from the GOES-R IR window channel data

window channel data

• calibrate the IR TBB data against co-located PMW retrievals calibrate the IR TBB data against co-located PMW retrievals

• We tested the prototype algorithm with the IR We tested the prototype algorithm with the IR (band-13) data from HIMAWARI-8 for October (band-13) data from HIMAWARI-8 for October 29, 2015

29, 2015

• data provided by UW data provided by UW

• 2kmx2km, 10min 2kmx2km, 10min

• This test demonstrated the feasibility and This test demonstrated the feasibility and improved performance of the

improved performance of the GOES-R GOES-R Precipitation Estimates (GPE)

Precipitation Estimates (GPE)

Preliminary Work on H8 IR Precip Estimates Preliminary Work on H8 IR Precip Estimates

[2] Prototype Algorithm Developed [2] Prototype Algorithm Developed

(14)

Preliminary Work on H8 IR Precip Estimates Preliminary Work on H8 IR Precip Estimates

[3] Sample Animation at 4km / 30-min over the Entire Domain [3] Sample Animation at 4km / 30-min over the Entire Domain

(15)

Preliminary Work on H8 IR Precip Estimates Preliminary Work on H8 IR Precip Estimates

[4] Sample Animation at 2km / 10-min over NW Pacific [4] Sample Animation at 2km / 10-min over NW Pacific

(16)

• The prototype algorithm performs quite wellThe prototype algorithm performs quite well

• More work needed to refine the details, More work needed to refine the details,

especially to improve the spatial continuity of the especially to improve the spatial continuity of the calibration tables

calibration tables

• Also need to work together with other GOES-R Also need to work together with other GOES-R PIs on improving the GPE with information from PIs on improving the GPE with information from other channels (water vapor, lighning et al.)

other channels (water vapor, lighning et al.)

• We are in process of infuse the GPE to our We are in process of infuse the GPE to our regional CMORPH system

regional CMORPH system

Preliminary Work on H8 IR Precip Estimates Preliminary Work on H8 IR Precip Estimates

[5] What We have Done and Learned [5] What We have Done and Learned

(17)

Summary Summary

• CPC has developed techniques to derived precipitation CPC has developed techniques to derived precipitation estimates from GEO and LEO IR data;

estimates from GEO and LEO IR data;

• The techniques are developed as part of the CMORPH The techniques are developed as part of the CMORPH work through PDF calibration against combined PMW work through PDF calibration against combined PMW retrievals and CloudSat Radar observations;

retrievals and CloudSat Radar observations;

• We also experimented to apply the PDF matching We also experimented to apply the PDF matching

technique to derived precipitation estimates from H8 high- technique to derived precipitation estimates from H8 high- resolution IR data;

resolution IR data;