Assimilation of CrIS and ATMS Radiances in HWRF to Improve Hurricane/Typhoon Forecasts

Assimilation of CrIS and ATMS Radiances in HWRF to Improve Hurricane/Typhoon Forecasts Fuzhong Weng 1, Xiaolei Zou 2, Lin Lin 3, Banglin Zhang 4, Vijay Tallaparagada 4 and Mitch Goldberg 5 1. NOAA Center for Satellite Applications and Research 2. University of Maryland 3. IMSG Group Inc. 4. NOAA/NCEP Environmental Modeling Center 5. NOAA JPSS Program Office 5 th Asia Oceania Meteorological Satellite User Conference, 19-21, 2014, Shanghai, China.

Outline Highlights on SNPP CrIS/ATMS CalVal Status Challenges of Satellite Radiance Assimilation in HWRF Approaches Developed for HWRF Radiance Assimilation Impacts of CRIS and ATMS on Hurricane Forecasts Summary and Future Works 2

Suomi NPP SDR Product Maturity Status Sensor Beta Provisional Validated CrIS February 10, 2012 February 6, 2013 March 18, 2014 ATMS May 2, 2012 February 12, 2013 March 18, 2014 C OMPS March 7, 2012 March 12, 2013 September 3, 2014 VIIRS May 2, 2012 March 13, 2013 April 16, 2014 C Beta C Early release product. Initial calibration applied Minimally validated and may still contain significant errors (rapid changes can be expected. Version changes will not be identified as errors are corrected as on-orbit baseline is not established) Available to allow users to gain familiarity it with data formats and parameters Product is not appropriate as the basis for quantitative scientific publications studies and applications Provisional Product quality may not be optimal Incremental product improvements are still occurring as calibration parameters are adjusted with sensor on-orbit characterization (versions will be tracked) General research community is encouraged to participate in the QA and validation of the product, but need to be aware that product validation and QA are ongoing Users are urged to consult the SDR product status document prior to use of the data in publications Ready for operational evaluation Validated On-orbit sensor performance characterized and calibration parameters adjusted accordingly Ready for use in applications and scientific publications There may be later improved versions There will be strong versioning with documentation 3

Advanced Technology Microwave Sounder (ATMS) TDR SDR Status ATMS SDR data have reached a full validated maturity level It has a stable instrument performance and calibration ATMS instrument noises (NEDT) are characterized and all channels meet or exceed the requirements All the ATMS channels have noises much lower than specification ATMS processing coefficients table (PCT) are updated with nominal values Quality flags (e.g. spacecraft maneuver and scanline, calibration) are updated and monitored Geolocation errors for all the channels are quantified and meet specification Instrument performance and SDR uncertainties are well characterized and documented (e.g. ATBD, peer-reviewed publications, user s guide, error budget analysis) Both TDR and SDR products are used in NOAA operations Innovated sciences have been made for instrument calibration 4

ATMS Channel Weighting Functions hpa) essure (h Pre Weighting Function 5

ATMS Noise Equivalent Differential Temperature (NEDT) NEΔT (K) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Channel Number On-orbit ATMS noise magnitudes are about twice as large as those AMSUMHS but much lower than specification. The re-sampled ATMS data within CrIS FOR or equivalent AMSU-A FOV would result in noise much lower than that of AMSU-A/MHS 6

Warm Core - Temperature Anomaly Typhoon Neoguri 07/07/2014

Cross-Track Infrared Sounder (CRIS) SDR Status CRIS SDR data have reached a full validated maturity level On-orbit instrument performance is stable and well within specification SDR uncertainties meet requirements, with NEdN and uncertainties of radiometric and spectral calibration well below specifications Spectral uncertainty: < 3 ppm, well below specification Radiometric uncertainty: ~0.1K, well below specification Geolocation error: < 1.2 km below specification ILS and NL correction algorithms/code and coefficients are improved and adjusted All the important SDR quality flags are functioning as expected 99.98% of the SDR data are in good quality IDPS capability to process full resolution RDRs is implemented and tested Instrument performance and SDR uncertainties are well characterized and documented (e.g. ATBD, peer-reviewed publications, user s guide, error budget analysis) CrIS SDR team developed lots of innovative algorithms for full resolution SDR processing 8

CRIS SDR Ingested in NCEP GFS (399channels) re (hpa) Pressur Assimilated in GFS and HWRF (84) 9

CrIS Noise Equivalent Differential Radiance (NEdN) Note: MWIR7 (Green) is known to have excess NEdN (expected) NEdN On Orbit data is consistent with Ground Test Data (Black Lines); Much Better Than Spec Limits (Dashed Lines) No Ice Contamination Signatures Slide courtesy of Joe Predina, ITT EXELIS 10

Challenges of Radiance Assimilation in HWRF Currently, satellite radiance assimilation is mostly conducted in the global NWP system where clear or non-precipitating radiances (microwave sounders) are assimilated NCEP uses Gridpoint Statistical Interpolation (GSI) system for both global and regional NWP. Some of key parameters in GSI such as observation error covariance and bias corrections are not fine tuned for the radiance assimilation in HWRF domains Hurricane Weather Research Forecast (HWRF) model is now raised to 2 hpa for HWRF data assimilation but this pressure level remains problematic for assimilating more upper air sounding channels HWRF data assimilation cycle is still cold-started and relies on the GFS analysis and 6- hour forecasts, thus requiring a bogus vortex constructed from the observations. The data used for constructing the vortex is very limited over vast oceans Assimilation of surface sensitive channels is always an issue if surface emitted radiation is poorly simulated from the forward model The earlier research from this project indicates the current GSI QC for microwave humidity sounders did not remove the radiances affected by clouds. Thus, large negative impact are seen on the precipitation forecast when MHS data are assimilated in the regional forecast model 12

New Approach for Direct Radiance Assimilation il i in HWRF Optimize the HWRF data assimilation system (e.g. higher model top and all domain assimilation) for effective ingestion and assimilation of satellite sounder radiances Examine the quality control procedures for all the ingested data and implement the additional criteria for removing all clouds-affected radiances from microwave humidity sounder Improve the radiative transfer models for better characterizing the error covariance within each of HWRF domains Refine the bias correction algorithms at various HWRF domains and according to cloud and precipitation type Conduct various observing system experiments (OSE) to understand the impacts of water vapor channels and surface sensitive channels 14

Weighting Functions for ATMS Channels Δp(hPa) Δp(hPa) L43 L61 Pressure (hp Pa) Ch. 15 Ch. 14 Ch. 13 Ch. 12 Ch. 11 Pressure (hp Pa) Weighting Function Weighting Function 15

HWRF Initialization and Data Assimilation Cold Start GFS ANL Initial Time Turn off GSI 5-day Forecast 0000 UTC 0000 UTC day5 GFS FST Turn off GSI 5-day Forecast 1800 UTC 0000 UTC 0000 UTC day5 Turn on GSI GFS ANL 5-day Forecast 0000 UTC 0000 UTC day5 GFS FST Turn on GSI 5-day Forecast 1800 UTC 0000 UTC 0000 UTC day5 Warm Start t HWRF FST Turn on GSI 5-day Forecast 1800 UTC 0000 UTC 0000 UTC day5 16

HWRF Model and Data Assimilation System HWRF Model: 2012 NCEP-Trunk version 934 Three telescoping domains: Outer domain: 27km: 75x75 o ; Inner domain: 9km ~11x10 o Inner-most domain: 3km inner-most nest ~6x6 66 o Revised Model Level and Top: Vertical llevels: l 61 Model top: 0.5 hpa Data Assimilation System: HWRF 6 hour forecasts GSI (3DVAR) The Hurricane Weather Research and Forecasting (HWRF) Model dynamical core is designed based on the WRF model using NCEP Non- Hydrostatic Mesoscale Model (NMM) core with a movable highresolution nested grid (telescopic) Regional-Scale, Moving Nest, Ocean-Atmosphere Coupled Modeling System. Horizontal resolution: 27 km outer grid, 9 km inner grid, 42 vertical levels Non-Hydrostaticsystem of equations formulated on a rotated latitudelongitude, Arakawa E-grid and a vertical, pressure hybrid (sigma_p-p) coordinate. Advanced HWRF 3D Variational analysis that includes vortex relocation, correction to winds, MSLP, temperature and moisture in the hurricane region and adjustment to actual storm intensity. Uses SAS convection scheme, GFS/GFDL surface, boundary layer physics, GFDL/GFS radiation and Ferrier Microphysical Scheme. 17 Ocean coupled modeling system (POM/HYCOM).

Observing System Experiments for Tropical Storm Dbb Debby June 23, 2012 Conventional only data (CONV) radiosondes, aircraft reports (AIREP/PIREP, RECCO, MDCRS-ACARS, TAMDAR, AMDAR), surface ship and buoy observations, surface observations over land, pibal winds, wind profilers, VAD wind, and dropsondes. Including satellite data from the following instruments (CONV+) AMSU-A (N18, N19, and MetOp-A), ATMS (SNPP) AIRS (Aqua), HIRS (N19 and MetOp-A), CrIS (SNPP) For all the sensors, no water vapor channels, no channels whose peak WF < 0.5 mb, and no surface sensitive channels (WF > 600 hpa) 18

Study Case: Tropical Storm Debby Operational predictions of Debby tracks are all wrong prior to June 24 Synoptic Conditions: The tropical storm Debby was located in between a subtropical trough located to its southeast and a middle latitude ridge located to its northwest. The anti-cyclonic flows on the west side of the subtropical high and on the east edge of the middle latitude ridge favored a cyclonic flow development and a low-pressure system in the Gulf of Mexico at 1800 UTC June 23, 2012. The middle latitude ridge experienced an enhanced development with time, preventing Debby s northwestward movement. The subtropical high gradually retreated eastwardly and the northeastward flow in the southwestward branch of the subtropical high and the middle latitude westerly drove Debby to move eastward on June 24 when approaching the Gulf coast. Debby made landfall in Florida on June 26. It continued its eastward movement and went across the Florida and moved into Atlantic Ocean. http://en.wikipedia.org/wiki/tropical_storm_debby_(20 12)#mediaviewer/File:Debby_2012_track.png GOES-13 visible image of Tropical Storm Debby on June 25, 2012 at 11:45 a.m. EDT.

Track Predictions of the 2012 Operational HWRF The operational HWRF model produces an westward propagating tracks while the actual track of Debby was northeastward when model forecasts were initialized before June 25, 2012. The operational HWRF model produces reasonably good track forecasts after June 25 and afterward. The track prediction of Debby before June 25, 2012 was a major challenge. 20

Observing System Experiments (1/2) CONV +AMSUA CONV+HIRS/4 CONV+ATMS CONV +AIRS CONV+CrIS

Observing System Experiments (2/2) CONV + AMSUA+ATMS CONV +AMSUA+CrIS CONV +AMSUA+ATMS+CrIS CONV +ATMS + CrIS It appears that assimilating the AMSU-A and ATMS data at the same time causes negative impact on the track forecast initialized at 1800 UTC 24 June, 2012.

Summary and Future Works HWRF/GSI model was re-configured to have more vertical layers and a higher model top for more effective assimilation of upper-level satellite sounding data Higher HWRF model top can generate an improved atmospheric steering flow and thus better predict the movement of tropical cyclones For Hurricane Debby in 2012, assimilation of ATMS and CrIS results in an overall positive impact s on hurricane forecast score, in comparison with other combinations In comparison with AMSU, ATMS has a better overall performance score. A combination of both instruments do not favor the track forecast. We need to make further diagnostics on how microwave sounders fight each other in HWRF data assimilation system 25

Backup Slides 26

2010-2012 Hurricane Average Forecast Errors Experimental results from a large sample of 3-season (2010-2011-2012) tests showed about 10-15% improvement in tracks and about 20-25% improvement in intensity forecasts H2FI: 2012 HWRF H3FI: 2013 HWRF H2FI: 2012 HWRF H3FI: 2013 HWRF Figures Courtesy: James Franklin, NHC 27